{ "meta": { "name": "Agent Ecosystem DB", "name_zh": "AI Agent生态知识引擎", "domain": "agent.genetech.tools", "description": "AI Agent ecosystem database covering MCP servers, SDKs, protocols, memory systems, and benchmarks", "updated": "2026-06-29T04:15:30.736Z", "total_entities": 574, "categories": [ "mcp_servers", "agent_sdks", "protocols", "memory_systems", "model_apis", "vector_dbs", "benchmarks", "sdks" ], "related_sites": [ { "domain": "robot.genetech.tools", "name": "RobotParts", "relation": "Agents control robots via protocols", "url": "https://robot.genetech.tools" }, { "domain": "brain.genetech.tools", "name": "BrainDB", "relation": "BCI-agent interfaces", "url": "https://brain.genetech.tools" }, { "domain": "quantum.genetech.tools", "name": "QuantumDB", "relation": "Quantum ML for agents", "url": "https://quantum.genetech.tools" } ] }, "data": { "mcp_servers": [ { "id": "MCP-001", "name": "server-filesystem", "full_name": "@modelcontextprotocol/server-filesystem", "category": "文件系统", "description": "读写本地文件系统", "protocol_version": "2025-03-26", "auth": "local", "hosts": [ "Claude Desktop", "VS Code", "Cursor" ], "security": "本地沙箱", "stars": 13000, "maintainer": "Anthropic", "status": "active", "language": "TypeScript", "features": [ "读写文件", "目录列表", "文件搜索" ] }, { "id": "MCP-002", "name": "server-postgres", "full_name": "@modelcontextprotocol/server-postgres", "category": "数据库", "description": "PostgreSQL 数据库查询", "protocol_version": "2025-03-26", "auth": "connection_string", "hosts": [ "Claude Desktop", "VS Code" ], "security": "只读查询模式", "stars": 8500, "maintainer": "Anthropic", "status": "active", "language": "TypeScript", "features": [ "SQL查询", "schema浏览", "只读模式" ] }, { "id": "MCP-003", "name": "server-github", "full_name": "@modelcontextprotocol/server-github", "category": "API", "description": "GitHub API 操作", "protocol_version": "2025-03-26", "auth": "OAuth token", "hosts": [ "Claude Desktop", "VS Code", "Cursor" ], "security": "token权限控制", "stars": 7200, "maintainer": "Anthropic", "status": "active", "language": "TypeScript", "features": [ "创建PR/Issue", "搜索代码", "文件操作" ] }, { "id": "MCP-004", "name": "server-puppeteer", "full_name": "@modelcontextprotocol/server-puppeteer", "category": "浏览器", "description": "浏览器自动化", "protocol_version": "2025-03-26", "auth": "none", "hosts": [ "Claude Desktop" ], "security": "沙箱浏览器", "stars": 6800, "maintainer": "Anthropic", "status": "active", "language": "TypeScript", "features": [ "网页截图", "表单填写", "点击交互" ] }, { "id": "MCP-005", "name": "server-sqlite", "full_name": "@modelcontextprotocol/server-sqlite", "category": "数据库", "description": "SQLite 本地数据库", "protocol_version": "2025-03-26", "auth": "none", "hosts": [ "Claude Desktop", "VS Code" ], "security": "本地文件", "stars": 5900, "maintainer": "Anthropic", "status": "active", "language": "TypeScript", "features": [ "SQL查询", "建表", "数据导入" ] }, { "id": "MCP-006", "name": "server-brave-search", "full_name": "@modelcontextprotocol/server-brave-search", "category": "搜索", "description": "Brave 搜索引擎", "protocol_version": "2025-03-26", "auth": "API key", "hosts": [ "Claude Desktop" ], "security": "API限流", "stars": 4500, "maintainer": "Anthropic", "status": "active", "language": "TypeScript", "features": [ "网页搜索", "结果过滤" ] }, { "id": "MCP-007", "name": "server-memory", "full_name": "@modelcontextprotocol/server-memory", "category": "记忆", "description": "Agent 持久化记忆", "protocol_version": "2025-03-26", "auth": "none", "hosts": [ "Claude Desktop" ], "security": "本地存储", "stars": 5200, "maintainer": "Anthropic", "status": "active", "language": "TypeScript", "features": [ "知识图谱", "实体关系", "持久化" ] }, { "id": "MCP-008", "name": "firecrawl-mcp-server", "full_name": "firecrawl-mcp-server", "category": "网页抓取", "description": "网页内容提取和爬取", "protocol_version": "2025-03-26", "auth": "API key", "hosts": [ "Claude Desktop", "Cursor" ], "security": "API限流", "stars": 3800, "maintainer": "Mendable", "status": "active", "language": "TypeScript", "features": [ "网页抓取", "批量爬取", "结构化提取" ] }, { "id": "MCP-009", "name": "server-sequential-thinking", "full_name": "@modelcontextprotocol/server-sequential-thinking", "category": "推理", "description": "结构化思维链推理", "protocol_version": "2025-03-26", "auth": "none", "hosts": [ "Claude Desktop" ], "security": "无风险", "stars": 4100, "maintainer": "Anthropic", "status": "active", "language": "TypeScript", "features": [ "思维链", "分步推理", "动态调整" ] }, { "id": "MCP-010", "name": "notion-mcp-server", "full_name": "@notionhq/notion-mcp-server", "category": "生产力", "description": "Notion 页面和数据库操作", "protocol_version": "2025-03-26", "auth": "OAuth", "hosts": [ "Claude Desktop", "Cursor" ], "security": "token权限", "stars": 3200, "maintainer": "Notion", "status": "active", "language": "TypeScript", "features": [ "页面CRUD", "数据库查询", "搜索" ] }, { "id": "MCP-011", "name": "server-slack", "full_name": "@modelcontextprotocol/server-slack", "category": "通信", "description": "Slack 消息和频道操作", "protocol_version": "2025-03-26", "auth": "Bot token", "hosts": [ "Claude Desktop" ], "security": "token权限", "stars": 2800, "maintainer": "Anthropic", "status": "active", "language": "TypeScript", "features": [ "发消息", "读频道", "搜索" ] }, { "id": "MCP-012", "name": "supabase-mcp-server", "full_name": "supabase-mcp-server", "category": "数据库", "description": "Supabase 数据库和认证", "protocol_version": "2025-03-26", "auth": "Service key", "hosts": [ "Claude Desktop", "Cursor" ], "security": "key权限", "stars": 2600, "maintainer": "Supabase", "status": "active", "language": "TypeScript", "features": [ "SQL查询", "Auth管理", "Storage操作" ] }, { "id": "MCP-013", "name": "Context7", "full_name": "@upstash/context7-mcp", "category": "文档", "description": "实时获取最新库文档和代码示例", "protocol_version": "2025-03-26", "auth": "none", "hosts": [ "Claude Desktop", "VS Code", "Cursor" ], "security": "只读", "stars": 4200, "maintainer": "Upstash", "status": "active", "language": "TypeScript", "features": [ "实时文档", "代码示例", "版本感知" ] }, { "id": "MCP-014", "name": "Exa MCP Server", "full_name": "exa-mcp-server", "category": "搜索", "description": "AI原生高级网络搜索引擎", "protocol_version": "2025-03-26", "auth": "API key", "hosts": [ "Claude Desktop", "Cursor" ], "security": "API限流", "stars": 3100, "maintainer": "Exa", "status": "active", "language": "TypeScript", "features": [ "语义搜索", "内容提取", "相似度搜索" ] }, { "id": "MCP-015", "name": "Browserbase", "full_name": "@browserbasehq/mcp-server", "category": "浏览器", "description": "云端浏览器自动化，支持多会话", "protocol_version": "2025-03-26", "auth": "API key", "hosts": [ "Claude Desktop", "Cursor" ], "security": "云端沙箱", "stars": 2800, "maintainer": "Browserbase", "status": "active", "language": "TypeScript", "features": [ "云端浏览器", "多会话", "截图/PDF" ] }, { "id": "MCP-016", "name": "server-docker", "full_name": "@modelcontextprotocol/server-docker", "category": "DevOps", "description": "Docker容器管理", "protocol_version": "2025-03-26", "auth": "local", "hosts": [ "Claude Desktop" ], "security": "Docker socket权限", "stars": 2400, "maintainer": "Anthropic", "status": "active", "language": "TypeScript", "features": [ "容器管理", "镜像操作", "日志查看" ] }, { "id": "MCP-017", "name": "server-git", "full_name": "@modelcontextprotocol/server-git", "category": "版本控制", "description": "Git仓库操作", "protocol_version": "2025-03-26", "auth": "local", "hosts": [ "Claude Desktop", "VS Code" ], "security": "本地权限", "stars": 2200, "maintainer": "Anthropic", "status": "active", "language": "TypeScript", "features": [ "Git操作", "diff查看", "分支管理" ] }, { "id": "MCP-018", "name": "server-google-maps", "full_name": "@modelcontextprotocol/server-google-maps", "category": "地图", "description": "Google Maps API集成", "protocol_version": "2025-03-26", "auth": "API key", "hosts": [ "Claude Desktop" ], "security": "API限流", "stars": 1800, "maintainer": "Anthropic", "status": "active", "language": "TypeScript", "features": [ "地理编码", "路线规划", "地点搜索" ] }, { "id": "MCP-019", "name": "server-fetch", "full_name": "@modelcontextprotocol/server-fetch", "category": "网络", "description": "HTTP请求和网页内容获取", "protocol_version": "2025-03-26", "auth": "none", "hosts": [ "Claude Desktop", "VS Code", "Cursor" ], "security": "沙箱请求", "stars": 3500, "maintainer": "Anthropic", "status": "active", "language": "TypeScript", "features": [ "HTTP GET/POST", "HTML转Markdown", "内容提取" ] }, { "id": "MCP-020", "name": "server-time", "full_name": "@modelcontextprotocol/server-time", "category": "工具", "description": "时间查询和时区转换", "protocol_version": "2025-03-26", "auth": "none", "hosts": [ "Claude Desktop" ], "security": "无风险", "stars": 1200, "maintainer": "Anthropic", "status": "active", "language": "TypeScript", "features": [ "当前时间", "时区转换", "时间计算" ] }, { "id": "MCP-021", "name": "Figma MCP Server", "full_name": "@figma/mcp-server", "category": "设计", "description": "Figma设计文件读取和代码生成", "protocol_version": "2025-03-26", "auth": "OAuth", "hosts": [ "Claude Desktop", "Cursor" ], "security": "token权限", "stars": 3600, "maintainer": "Figma", "status": "active", "language": "TypeScript", "features": [ "设计读取", "代码生成", "布局分析" ] }, { "id": "MCP-022", "name": "Linear MCP Server", "full_name": "linear-mcp-server", "category": "项目管理", "description": "Linear项目管理工具集成", "protocol_version": "2025-03-26", "auth": "API key", "hosts": [ "Claude Desktop", "Cursor" ], "security": "API权限", "stars": 1900, "maintainer": "Linear", "status": "active", "language": "TypeScript", "features": [ "Issue管理", "项目追踪", "团队协作" ] }, { "id": "MCP-023", "name": "Jira MCP Server", "full_name": "jira-mcp-server", "category": "项目管理", "description": "Jira项目管理集成", "protocol_version": "2025-03-26", "auth": "OAuth/API token", "hosts": [ "Claude Desktop" ], "security": "token权限", "stars": 1500, "maintainer": "社区", "status": "active", "language": "TypeScript", "features": [ "Issue CRUD", "Sprint管理", "JQL查询" ] }, { "id": "MCP-024", "name": "server-sentry", "full_name": "@modelcontextprotocol/server-sentry", "category": "监控", "description": "Sentry错误追踪集成", "protocol_version": "2025-03-26", "auth": "API token", "hosts": [ "Claude Desktop" ], "security": "token权限", "stars": 1400, "maintainer": "Sentry", "status": "active", "language": "TypeScript", "features": [ "错误查看", "性能监控", "Release追踪" ] }, { "id": "MCP-025", "name": "Stripe MCP Server", "full_name": "stripe-mcp-server", "category": "支付", "description": "Stripe支付平台集成", "protocol_version": "2025-03-26", "auth": "API key", "hosts": [ "Claude Desktop" ], "security": "受限key", "stars": 1600, "maintainer": "Stripe", "status": "active", "language": "TypeScript", "features": [ "客户管理", "订阅管理", "发票操作" ] }, { "id": "MCP-026", "name": "Shopify MCP Server", "full_name": "shopify-mcp-server", "category": "电商", "description": "Shopify电商平台集成", "protocol_version": "2025-03-26", "auth": "OAuth", "hosts": [ "Claude Desktop" ], "security": "OAuth权限", "stars": 1100, "maintainer": "社区", "status": "active", "language": "TypeScript", "features": [ "产品管理", "订单管理", "库存查询" ] }, { "id": "MCP-027", "name": "MongoDB MCP Server", "full_name": "mongodb-mcp-server", "category": "数据库", "description": "MongoDB数据库操作", "protocol_version": "2025-03-26", "auth": "connection_string", "hosts": [ "Claude Desktop", "VS Code" ], "security": "连接权限", "stars": 2100, "maintainer": "MongoDB", "status": "active", "language": "TypeScript", "features": [ "CRUD操作", "聚合查询", "索引管理" ] }, { "id": "MCP-028", "name": "Redis MCP Server", "full_name": "redis-mcp-server", "category": "数据库", "description": "Redis缓存和消息队列操作", "protocol_version": "2025-03-26", "auth": "connection_string", "hosts": [ "Claude Desktop" ], "security": "连接权限", "stars": 1700, "maintainer": "社区", "status": "active", "language": "TypeScript", "features": [ "键值操作", "发布订阅", "缓存管理" ] }, { "id": "MCP-029", "name": "Anki MCP Server", "full_name": "anki-mcp-server", "category": "教育", "description": "Anki间隔重复闪卡应用集成", "protocol_version": "2025-03-26", "auth": "local", "hosts": [ "Claude Desktop" ], "security": "本地API", "stars": 900, "maintainer": "社区", "status": "active", "language": "TypeScript", "features": [ "创建卡片", "查询卡片", "学习进度" ] }, { "id": "MCP-030", "name": "server-google-drive", "full_name": "@modelcontextprotocol/server-google-drive", "category": "文件系统", "description": "Google Drive文件操作", "protocol_version": "2025-03-26", "auth": "OAuth", "hosts": [ "Claude Desktop" ], "security": "OAuth权限", "stars": 2000, "maintainer": "Anthropic", "status": "active", "language": "TypeScript", "features": [ "文件搜索", "读写文档", "列表查看" ] }, { "id": "MCP-031", "name": "server-aws", "full_name": "aws-mcp-server", "category": "云服务", "description": "AWS服务集成（S3/Lambda/DynamoDB等）", "protocol_version": "2025-03-26", "auth": "IAM credentials", "hosts": [ "Claude Desktop" ], "security": "IAM权限", "stars": 2300, "maintainer": "AWS", "status": "active", "language": "TypeScript", "features": [ "S3操作", "Lambda管理", "DynamoDB查询" ] }, { "id": "MCP-032", "name": "server-kubernetes", "full_name": "kubernetes-mcp-server", "category": "DevOps", "description": "Kubernetes集群管理", "protocol_version": "2025-03-26", "auth": "kubeconfig", "hosts": [ "Claude Desktop" ], "security": "RBAC权限", "stars": 1800, "maintainer": "社区", "status": "active", "language": "TypeScript", "features": [ "Pod管理", "部署操作", "日志查看" ] }, { "id": "MCP-033", "name": "Terraform MCP Server", "full_name": "terraform-mcp-server", "category": "DevOps", "description": "Terraform基础设施即代码", "protocol_version": "2025-03-26", "auth": "local", "hosts": [ "Claude Desktop" ], "security": "本地权限", "stars": 1300, "maintainer": "HashiCorp", "status": "active", "language": "TypeScript", "features": [ "Plan/Apply", "State查看", "模块浏览" ] }, { "id": "MCP-034", "name": "Playwright MCP Server", "full_name": "@playwright/mcp-server", "category": "浏览器", "description": "Playwright浏览器自动化，支持多浏览器", "protocol_version": "2025-03-26", "auth": "none", "hosts": [ "Claude Desktop", "VS Code", "Cursor" ], "security": "沙箱浏览器", "stars": 3200, "maintainer": "Microsoft", "status": "active", "language": "TypeScript", "features": [ "多浏览器", "截图/PDF", "网络拦截" ] }, { "id": "MCP-035", "name": "server-echosign", "full_name": "adobe-mcp-server", "category": "文档", "description": "Adobe文档服务集成（PDF操作等）", "protocol_version": "2025-03-26", "auth": "OAuth", "hosts": [ "Claude Desktop" ], "security": "OAuth权限", "stars": 800, "maintainer": "Adobe", "status": "active", "language": "TypeScript", "features": [ "PDF操作", "签名管理", "文档转换" ] }, { "id": "MCP-036", "name": "Sysdig MCP Server", "full_name": "sysdig-mcp-server", "category": "安全", "description": "云安全监控和威胁检测", "protocol_version": "2025-03-26", "auth": "API token", "hosts": [ "Claude Desktop" ], "security": "token权限", "stars": 700, "maintainer": "Sysdig", "status": "active", "language": "TypeScript", "features": [ "安全事件", "合规检查", "漏洞扫描" ] }, { "id": "MCP-037", "name": "server-twitter", "full_name": "twitter-mcp-server", "category": "社交媒体", "description": "Twitter/X平台操作", "protocol_version": "2025-03-26", "auth": "OAuth 2.0", "hosts": [ "Claude Desktop" ], "security": "OAuth权限", "stars": 1500, "maintainer": "社区", "status": "active", "language": "TypeScript", "features": [ "发推文", "搜索推文", "用户信息" ] }, { "id": "MCP-038", "name": "server-discord", "full_name": "discord-mcp-server", "category": "通信", "description": "Discord消息和频道操作", "protocol_version": "2025-03-26", "auth": "Bot token", "hosts": [ "Claude Desktop" ], "security": "Bot权限", "stars": 1200, "maintainer": "社区", "status": "active", "language": "TypeScript", "features": [ "发消息", "读频道", "服务器管理" ] }, { "id": "MCP-039", "name": "server-email (IMAP/SMTP)", "full_name": "email-mcp-server", "category": "通信", "description": "邮件收发操作（IMAP/SMTP）", "protocol_version": "2025-03-26", "auth": "credentials", "hosts": [ "Claude Desktop" ], "security": "加密连接", "stars": 1100, "maintainer": "社区", "status": "active", "language": "TypeScript", "features": [ "收邮件", "发邮件", "搜索邮件" ] }, { "id": "MCP-040", "name": "server-calendar (Google Calendar)", "full_name": "google-calendar-mcp-server", "category": "生产力", "description": "Google日历操作", "protocol_version": "2025-03-26", "auth": "OAuth", "hosts": [ "Claude Desktop" ], "security": "OAuth权限", "stars": 1600, "maintainer": "社区", "status": "active", "language": "TypeScript", "features": [ "创建事件", "查询日程", "冲突检测" ] }, { "id": "MCP-041", "name": "server-airtable", "full_name": "airtable-mcp-server", "category": "数据库", "description": "Airtable数据库操作", "protocol_version": "2025-03-26", "auth": "API key", "hosts": [ "Claude Desktop" ], "security": "API权限", "stars": 1000, "maintainer": "社区", "status": "active", "language": "TypeScript", "features": [ "记录CRUD", "视图查询", "字段操作" ] }, { "id": "MCP-042", "name": "server-snowflake", "full_name": "snowflake-mcp-server", "category": "数据库", "description": "Snowflake数据仓库查询", "protocol_version": "2025-03-26", "auth": "OAuth/key", "hosts": [ "Claude Desktop" ], "security": "RBAC权限", "stars": 900, "maintainer": "Snowflake", "status": "active", "language": "TypeScript", "features": [ "SQL查询", "数据仓库", "分析函数" ] }, { "id": "MCP-043", "name": "server-elasticsearch", "full_name": "elasticsearch-mcp-server", "category": "搜索", "description": "Elasticsearch全文搜索集成", "protocol_version": "2025-03-26", "auth": "API key", "hosts": [ "Claude Desktop" ], "security": "API权限", "stars": 1300, "maintainer": "Elastic", "status": "active", "language": "TypeScript", "features": [ "全文搜索", "聚合分析", "索引管理" ] }, { "id": "MCP-044", "name": "server-langchain", "full_name": "langchain-mcp-server", "category": "AI", "description": "LangChain工具和链集成", "protocol_version": "2025-03-26", "auth": "varies", "hosts": [ "Claude Desktop" ], "security": "工具权限", "stars": 1800, "maintainer": "LangChain", "status": "active", "language": "TypeScript", "features": [ "链调用", "RAG管道", "工具集成" ] }, { "id": "MCP-045", "name": "server-computer-use", "full_name": "@anthropic/server-computer-use", "category": "桌面自动化", "description": "桌面GUI自动化操作", "protocol_version": "2025-03-26", "auth": "local", "hosts": [ "Claude Desktop" ], "security": "本地沙箱", "stars": 2900, "maintainer": "Anthropic", "status": "active", "language": "TypeScript", "features": [ "屏幕截图", "鼠标键盘", "应用操作" ] }, { "id": "MCP-046", "name": "Vercel MCP Server", "full_name": "vercel-mcp-server", "category": "部署", "description": "Vercel部署平台集成", "protocol_version": "2025-03-26", "auth": "API token", "hosts": [ "Claude Desktop", "Cursor" ], "security": "token权限", "stars": 1700, "maintainer": "Vercel", "status": "active", "language": "TypeScript", "features": [ "部署管理", "域名操作", "日志查看" ] }, { "id": "MCP-047", "name": "Neon MCP Server", "full_name": "neon-mcp-server", "category": "数据库", "description": "Neon Serverless Postgres操作", "protocol_version": "2025-03-26", "auth": "API key", "hosts": [ "Claude Desktop", "Cursor" ], "security": "API权限", "stars": 1100, "maintainer": "Neon", "status": "active", "language": "TypeScript", "features": [ "分支管理", "SQL查询", "自动扩缩" ] }, { "id": "MCP-048", "name": "Cloudflare MCP Server", "full_name": "cloudflare-mcp-server", "category": "云服务", "description": "Cloudflare Workers和DNS管理", "protocol_version": "2025-03-26", "auth": "API token", "hosts": [ "Claude Desktop" ], "security": "token权限", "stars": 1400, "maintainer": "Cloudflare", "status": "active", "language": "TypeScript", "features": [ "Workers部署", "DNS管理", "缓存清除" ] }, { "id": "MCP-049", "name": "server-huggingface", "full_name": "huggingface-mcp-server", "category": "AI", "description": "HuggingFace模型和数据集操作", "protocol_version": "2025-03-26", "auth": "API token", "hosts": [ "Claude Desktop" ], "security": "token权限", "stars": 1500, "maintainer": "HuggingFace", "status": "active", "language": "TypeScript", "features": [ "模型搜索", "推理调用", "数据集浏览" ] }, { "id": "MCP-050", "name": "server-openapi", "full_name": "@modelcontextprotocol/server-openapi", "category": "API", "description": "OpenAPI规范自动生成MCP工具", "protocol_version": "2025-03-26", "auth": "varies", "hosts": [ "Claude Desktop" ], "security": "规范定义", "stars": 1900, "maintainer": "Anthropic", "status": "active", "language": "TypeScript", "features": [ "自动工具生成", "API发现", "参数验证" ] }, { "id": "MCP-051", "name": "server-mysql", "full_name": "mysql-mcp-server", "category": "数据库", "description": "MySQL数据库查询操作", "protocol_version": "2025-03-26", "auth": "connection_string", "hosts": [ "Claude Desktop", "VS Code" ], "security": "连接权限", "stars": 1600, "maintainer": "社区", "status": "active", "language": "TypeScript", "features": [ "SQL查询", "Schema浏览", "数据导入导出" ] }, { "id": "MCP-052", "name": "server-zapier", "full_name": "zapier-mcp-server", "category": "自动化", "description": "Zapier工作流自动化集成", "protocol_version": "2025-03-26", "auth": "API key", "hosts": [ "Claude Desktop" ], "security": "API权限", "stars": 1300, "maintainer": "Zapier", "status": "active", "language": "TypeScript", "features": [ "触发器管理", "动作执行", "工作流编排" ] }, { "id": "MCP-google-cloud-db", "name": "Google Cloud Database MCP Servers", "type": "database", "company": "Google Cloud", "key_features": [ "Managed remote MCP for AlloyDB PostgreSQL, Spanner, Cloud SQL, Bigtable, Firestore", "Enterprise-grade auditing, observability and governance", "Natural language database interaction", "Schema creation, query diagnosis, vector similarity search" ], "status": "production", "year": 2026, "sources": [ { "source_type": "web", "source_credibility": "A", "collected_at": "2026-05-29T14:25:59.103Z" } ] }, { "id": "MCP-oracle-sqlcl", "name": "Oracle SQLcl MCP Server", "type": "database", "company": "Oracle", "key_features": [ "MCP server for Oracle Database via SQLcl", "AI agents can connect, run SQL/PLSQL queries and scripts", "Credentials managed locally on user machine", "All interactions logged in DBTOOLS$MCP_LOG", "VS Code SQL Developer integration with auto-registration" ], "status": "production", "year": 2025, "sources": [ { "source_type": "web", "source_credibility": "A", "collected_at": "2026-05-29T14:25:59.103Z" } ] }, { "id": "MCP-database-mcp", "name": "@cocaxcode/database-mcp", "type": "database", "company": "Community", "key_features": [ "33 tools across 3 engines (PostgreSQL, MySQL, SQLite)", "Connection groups, named connection management", "Automatic rollback with reverse SQL", "Schema auto-discovery via MCP Resources", "Zero config, 100% local credentials", "Works with Claude Code, Cursor, Windsurf, VS Code, Gemini CLI" ], "status": "production", "year": 2026, "sources": [ { "source_type": "web", "source_credibility": "B", "collected_at": "2026-05-29T14:25:59.103Z" } ] }, { "id": "MCP-universal-db", "name": "universal-db-mcp", "type": "database", "company": "Community (Anarkh-Lee)", "key_features": [ "Supports 17 databases: MySQL, PostgreSQL, Oracle, MongoDB, etc.", "Natural language query and analysis", "Compatible with 50+ platforms (Claude Desktop, Cursor, VS Code, ChatGPT)", "44 releases, active development" ], "status": "production", "year": 2026, "sources": [ { "source_type": "web", "source_credibility": "B", "collected_at": "2026-05-29T14:25:59.103Z" } ] }, { "id": "MCP-data-commons", "name": "Google Data Commons MCP Server", "type": "data_analysis", "company": "Google", "key_features": [ "Access to Google Data Commons public datasets", "Statistical data about populations, economics, environment", "Seamless integration with Google Cloud ecosystem" ], "status": "production", "year": 2025, "sources": [ { "source_type": "web", "source_credibility": "A", "collected_at": "2026-05-29T14:25:59.103Z" } ] }, { "id": "MCP-truefoundry-gateway", "name": "TrueFoundry MCP Gateway", "type": "gateway", "company": "TrueFoundry", "key_features": [ "3-4ms latency (10ms under load)", "350+ requests per second", "Enterprise-grade MCP gateway for AI agent security", "Centralized tool management and access control" ], "status": "production", "year": 2026, "sources": [ { "source_type": "web", "source_credibility": "B", "collected_at": "2026-05-29T14:25:59.103Z" } ] }, { "id": "MCP-skyvia", "name": "Skyvia MCP Server", "full_name": "Skyvia MCP", "category": "数据集成", "description": "云端数据集成平台MCP Server，支持ETL/ELT数据管道", "protocol_version": "2025-11-25", "auth": "API Key", "hosts": [ "Claude Desktop", "Cursor" ], "security": "OAuth 2.0", "stars": 800, "maintainer": "Skyvia", "status": "active", "language": "TypeScript", "features": [ "数据集成", "ETL管道", "云数据库连接" ], "sources": [ { "source_type": "web", "source_credibility": "B", "collected_at": "2026-05-30T13:38:30.808Z" } ] }, { "id": "MCP-ahrefs", "name": "Ahrefs MCP Server", "full_name": "ahrefs-mcp", "category": "SEO", "description": "SEO分析工具MCP Server，提供关键词研究、反向链接分析", "protocol_version": "2025-11-25", "auth": "API Key", "hosts": [ "Claude Desktop", "Cursor" ], "security": "API Key", "stars": 500, "maintainer": "Ahrefs", "status": "active", "language": "TypeScript", "features": [ "关键词研究", "反向链接分析", "SEO审计" ], "sources": [ { "source_type": "web", "source_credibility": "B", "collected_at": "2026-05-30T13:38:30.808Z" } ] }, { "id": "MCP-mikrotik", "name": "MikroTik Router MCP Server", "full_name": "io.github.kivanccakmak/yaver", "category": "网络管理", "description": "MikroTik路由器管理MCP Server，支持防火墙、NAT、DHCP、DNS、WireGuard", "protocol_version": "2025-11-25", "auth": "SSH", "hosts": [ "Claude Desktop" ], "security": "SSH密钥", "stars": 200, "maintainer": "社区", "status": "active", "language": "Python", "features": [ "防火墙管理", "NAT配置", "WireGuard VPN", "DHCP/DNS" ], "sources": [ { "source_type": "web", "source_credibility": "B", "collected_at": "2026-05-30T13:38:30.808Z" } ] }, { "id": "MCP-fastmcp", "name": "FastMCP Server", "full_name": "@fastmcp/server", "category": "开发框架", "description": "高性能MCP Server开发框架，用于快速构建自定义MCP Server", "protocol_version": "2025-11-25", "auth": "local", "hosts": [ "Claude Desktop", "Cursor", "VS Code" ], "security": "本地沙箱", "stars": 3500, "maintainer": "FastMCP", "status": "active", "language": "Python", "features": [ "快速开发", "热重载", "类型安全", "自动文档" ], "sources": [ { "source_type": "web", "source_credibility": "A", "collected_at": "2026-05-30T13:38:30.808Z" } ] }, { "id": "MCP-vercel-v0", "name": "Vercel v0 MCP Server", "full_name": "@vercel/v0-mcp", "category": "前端开发", "description": "Vercel v0 AI前端生成工具MCP Server", "protocol_version": "2025-11-25", "auth": "OAuth", "hosts": [ "Claude Desktop", "Cursor" ], "security": "OAuth 2.0", "stars": 2800, "maintainer": "Vercel", "status": "active", "language": "TypeScript", "features": [ "AI前端生成", "组件创建", "代码生成" ], "sources": [ { "source_type": "web", "source_credibility": "A", "collected_at": "2026-05-30T13:38:30.808Z" } ] }, { "id": "MCP-razorpay", "name": "Razorpay MCP Server", "full_name": "razorpay-mcp", "category": "支付", "description": "印度支付平台Razorpay的MCP Server", "protocol_version": "2025-11-25", "auth": "API Key", "hosts": [ "Claude Desktop" ], "security": "API Key", "stars": 150, "maintainer": "Razorpay", "status": "active", "language": "TypeScript", "features": [ "支付处理", "退款管理", "订阅管理" ], "sources": [ { "source_type": "web", "source_credibility": "B", "collected_at": "2026-05-30T13:38:30.808Z" } ] }, { "id": "MCP-linear-v2", "name": "Linear MCP Server v2", "full_name": "@linear/mcp-server", "category": "项目管理", "description": "Linear项目管理工具MCP Server升级版，支持更丰富的查询和操作", "protocol_version": "2025-11-25", "auth": "OAuth", "hosts": [ "Claude Desktop", "Cursor", "VS Code" ], "security": "OAuth 2.0", "stars": 1200, "maintainer": "Linear", "status": "active", "language": "TypeScript", "features": [ "项目管理", "Issue追踪", "团队协作", "工作流自动化" ], "sources": [ { "source_type": "web", "source_credibility": "A", "collected_at": "2026-05-30T13:38:30.808Z" } ] }, { "id": "MCP-1password", "name": "1Password MCP Server", "full_name": "@1password/mcp-server", "category": "安全", "description": "1Password密码管理器MCP Server，安全访问密钥和凭证", "protocol_version": "2025-11-25", "auth": "OAuth", "hosts": [ "Claude Desktop" ], "security": "端到端加密", "stars": 900, "maintainer": "1Password", "status": "active", "language": "TypeScript", "features": [ "密钥管理", "凭证访问", "安全存储" ], "sources": [ { "source_type": "web", "source_credibility": "A", "collected_at": "2026-05-30T13:38:30.808Z" } ] }, { "id": "MCP-hubspot", "name": "HubSpot MCP Server", "full_name": "hubspot-mcp", "category": "CRM", "description": "HubSpot CRM平台MCP Server", "protocol_version": "2025-11-25", "auth": "OAuth", "hosts": [ "Claude Desktop", "Cursor" ], "security": "OAuth 2.0", "stars": 600, "maintainer": "社区", "status": "active", "language": "TypeScript", "features": [ "CRM操作", "联系人管理", "交易追踪" ], "sources": [ { "source_type": "web", "source_credibility": "B", "collected_at": "2026-05-30T13:38:30.808Z" } ] }, { "id": "MCP-resend", "name": "Resend MCP Server", "full_name": "resend-mcp", "category": "邮件", "description": "Resend邮件服务MCP Server，开发者友好的邮件API", "protocol_version": "2025-11-25", "auth": "API Key", "hosts": [ "Claude Desktop", "Cursor" ], "security": "API Key", "stars": 400, "maintainer": "Resend", "status": "active", "language": "TypeScript", "features": [ "邮件发送", "模板管理", "邮件追踪" ], "sources": [ { "source_type": "web", "source_credibility": "B", "collected_at": "2026-05-30T13:38:30.808Z" } ] }, { "id": "MCP-sentry-v2", "name": "Sentry MCP Server v2", "full_name": "@sentry/mcp-server", "category": "监控", "description": "Sentry错误监控MCP Server升级版，支持AI辅助错误分析", "protocol_version": "2025-11-25", "auth": "API Key", "hosts": [ "Claude Desktop", "Cursor" ], "security": "API Key", "stars": 1100, "maintainer": "Sentry", "status": "active", "language": "TypeScript", "features": [ "错误追踪", "性能监控", "AI错误分析", "Release追踪" ], "sources": [ { "source_type": "web", "source_credibility": "A", "collected_at": "2026-05-30T13:38:30.808Z" } ] }, { "id": "MCP-postmark", "name": "Postmark MCP Server", "full_name": "postmark-mcp", "category": "邮件", "description": "Postmark邮件投递服务MCP Server", "protocol_version": "2025-11-25", "auth": "API Key", "hosts": [ "Claude Desktop" ], "security": "API Key", "stars": 200, "maintainer": "社区", "status": "active", "language": "TypeScript", "features": [ "邮件投递", "退信处理", "邮件统计" ], "sources": [ { "source_type": "web", "source_credibility": "B", "collected_at": "2026-05-30T13:38:30.808Z" } ] }, { "id": "MCP-gdrive-v2", "name": "Google Drive MCP Server v2", "full_name": "@google/drive-mcp", "category": "文件系统", "description": "Google Drive MCP Server升级版，支持文件搜索、读写和共享", "protocol_version": "2025-11-25", "auth": "OAuth", "hosts": [ "Claude Desktop", "Cursor" ], "security": "OAuth 2.0", "stars": 2200, "maintainer": "Google", "status": "active", "language": "TypeScript", "features": [ "文件读写", "搜索", "共享管理", "版本控制" ], "sources": [ { "source_type": "web", "source_credibility": "A", "collected_at": "2026-05-30T13:38:30.808Z" } ] }, { "id": "MCP-clickup", "name": "ClickUp MCP Server", "full_name": "clickup-mcp", "category": "项目管理", "description": "ClickUp项目管理平台MCP Server", "protocol_version": "2025-11-25", "auth": "OAuth", "hosts": [ "Claude Desktop", "Cursor" ], "security": "OAuth 2.0", "stars": 350, "maintainer": "社区", "status": "active", "language": "TypeScript", "features": [ "任务管理", "文档协作", "时间追踪" ], "sources": [ { "source_type": "web", "source_credibility": "B", "collected_at": "2026-05-30T13:38:30.808Z" } ] }, { "id": "MCP-intercom", "name": "Intercom MCP Server", "full_name": "intercom-mcp", "category": "客服", "description": "Intercom客户服务平台MCP Server", "protocol_version": "2025-11-25", "auth": "OAuth", "hosts": [ "Claude Desktop" ], "security": "OAuth 2.0", "stars": 280, "maintainer": "社区", "status": "active", "language": "TypeScript", "features": [ "客户支持", "对话管理", "用户数据" ], "sources": [ { "source_type": "web", "source_credibility": "B", "collected_at": "2026-05-30T13:38:30.808Z" } ] }, { "id": "MCP-woocommerce", "name": "WooCommerce MCP Server", "full_name": "woocommerce-mcp", "category": "电商", "description": "WooCommerce电商平台MCP Server", "protocol_version": "2025-11-25", "auth": "API Key", "hosts": [ "Claude Desktop" ], "security": "API Key + HTTPS", "stars": 300, "maintainer": "社区", "status": "active", "language": "PHP/TypeScript", "features": [ "产品管理", "订单处理", "库存管理" ], "sources": [ { "source_type": "web", "source_credibility": "B", "collected_at": "2026-05-30T13:38:30.808Z" } ] }, { "id": "MCP-bigquery", "name": "BigQuery MCP Server", "full_name": "@google/bigquery-mcp", "category": "数据库", "description": "Google BigQuery数据仓库MCP Server", "protocol_version": "2025-11-25", "auth": "Service Account", "hosts": [ "Claude Desktop", "Cursor" ], "security": "IAM", "stars": 1800, "maintainer": "Google", "status": "active", "language": "TypeScript", "features": [ "SQL查询", "数据分析", "数据导出" ], "sources": [ { "source_type": "web", "source_credibility": "A", "collected_at": "2026-05-30T13:38:30.808Z" } ] }, { "id": "MCP-datadog", "name": "Datadog MCP Server", "full_name": "datadog-mcp", "category": "监控", "description": "Datadog监控平台MCP Server", "protocol_version": "2025-11-25", "auth": "API Key", "hosts": [ "Claude Desktop", "Cursor" ], "security": "API Key + APP Key", "stars": 700, "maintainer": "Datadog", "status": "active", "language": "TypeScript", "features": [ "指标查询", "日志搜索", "告警管理", "仪表盘" ], "sources": [ { "source_type": "web", "source_credibility": "A", "collected_at": "2026-05-30T13:38:30.808Z" } ] }, { "id": "MCP-twilio", "name": "Twilio MCP Server", "full_name": "twilio-mcp", "category": "通信", "description": "Twilio通信平台MCP Server，支持SMS、语音、视频", "protocol_version": "2025-11-25", "auth": "API Key", "hosts": [ "Claude Desktop", "Cursor" ], "security": "API Key", "stars": 550, "maintainer": "Twilio", "status": "active", "language": "TypeScript", "features": [ "SMS发送", "语音通话", "视频会议", "号码管理" ], "sources": [ { "source_type": "web", "source_credibility": "A", "collected_at": "2026-05-30T13:38:30.808Z" } ] }, { "id": "MCP-asana", "name": "Asana MCP Server", "full_name": "asana-mcp", "category": "项目管理", "description": "Asana项目管理MCP Server", "protocol_version": "2025-11-25", "auth": "OAuth", "hosts": [ "Claude Desktop", "Cursor" ], "security": "OAuth 2.0", "stars": 400, "maintainer": "社区", "status": "active", "language": "TypeScript", "features": [ "任务管理", "项目追踪", "团队协作" ], "sources": [ { "source_type": "web", "source_credibility": "B", "collected_at": "2026-05-30T13:38:30.808Z" } ] }, { "id": "MCP-figma-v2", "name": "Figma MCP Server v2", "full_name": "@figma/mcp-server-v2", "category": "设计", "description": "Figma设计工具MCP Server升级版，支持设计稿读取和代码生成", "protocol_version": "2025-11-25", "auth": "OAuth", "hosts": [ "Claude Desktop", "Cursor", "VS Code" ], "security": "OAuth 2.0", "stars": 3200, "maintainer": "Figma", "status": "active", "language": "TypeScript", "features": [ "设计稿读取", "代码生成", "组件库访问", "样式提取" ], "sources": [ { "source_type": "web", "source_credibility": "A", "collected_at": "2026-05-30T13:38:30.808Z" } ] }, { "id": "MCP-xero", "name": "Xero MCP Server", "full_name": "xero-mcp", "category": "财务", "description": "Xero财务软件MCP Server", "protocol_version": "2025-11-25", "auth": "OAuth", "hosts": [ "Claude Desktop" ], "security": "OAuth 2.0", "stars": 200, "maintainer": "社区", "status": "active", "language": "TypeScript", "features": [ "发票管理", "银行对账", "财务报告" ], "sources": [ { "source_type": "web", "source_credibility": "B", "collected_at": "2026-05-30T13:38:30.808Z" } ] }, { "id": "MCP-quickbooks", "name": "QuickBooks MCP Server", "full_name": "quickbooks-mcp", "category": "财务", "description": "QuickBooks会计软件MCP Server", "protocol_version": "2025-11-25", "auth": "OAuth", "hosts": [ "Claude Desktop" ], "security": "OAuth 2.0", "stars": 250, "maintainer": "社区", "status": "active", "language": "TypeScript", "features": [ "会计管理", "发票处理", "薪资管理" ], "sources": [ { "source_type": "web", "source_credibility": "B", "collected_at": "2026-05-30T13:38:30.808Z" } ] }, { "id": "MCP-miro", "name": "Miro MCP Server", "full_name": "miro-mcp", "category": "协作", "description": "Miro在线白板协作平台MCP Server", "protocol_version": "2025-11-25", "auth": "OAuth", "hosts": [ "Claude Desktop", "Cursor" ], "security": "OAuth 2.0", "stars": 350, "maintainer": "社区", "status": "active", "language": "TypeScript", "features": [ "白板操作", "协作绘图", "模板管理" ], "sources": [ { "source_type": "web", "source_credibility": "B", "collected_at": "2026-05-30T13:38:30.808Z" } ] }, { "id": "MCP-confluence", "name": "Confluence MCP Server", "full_name": "confluence-mcp", "category": "知识管理", "description": "Atlassian Confluence知识库MCP Server", "protocol_version": "2025-11-25", "auth": "OAuth", "hosts": [ "Claude Desktop", "Cursor" ], "security": "OAuth 2.0", "stars": 800, "maintainer": "社区", "status": "active", "language": "TypeScript", "features": [ "页面读写", "空间管理", "搜索", "评论" ], "sources": [ { "source_type": "web", "source_credibility": "B", "collected_at": "2026-05-30T13:38:30.808Z" } ] }, { "id": "MCP-github-v2", "name": "GitHub MCP Server v2", "full_name": "@github/mcp-server-v2", "category": "开发", "description": "GitHub MCP Server升级版，支持Copilot集成和更丰富的API", "protocol_version": "2025-11-25", "auth": "OAuth", "hosts": [ "Claude Desktop", "Cursor", "VS Code" ], "security": "OAuth 2.0", "stars": 8000, "maintainer": "GitHub", "status": "active", "language": "TypeScript", "features": [ "PR管理", "Issue操作", "Copilot集成", "Actions触发", "代码搜索" ], "sources": [ { "source_type": "web", "source_credibility": "A", "collected_at": "2026-05-30T13:38:30.808Z" } ] }, { "id": "MCP-mcpregistry", "name": "MCP Registry Server", "full_name": "@mcp/registry", "category": "注册中心", "description": "MCP官方注册中心Server，用于发现和验证MCP Server", "protocol_version": "2025-11-25", "auth": "local", "hosts": [ "Claude Desktop", "Cursor" ], "security": "TLS", "stars": 1500, "maintainer": "Linux Foundation/MCP", "status": "active", "language": "TypeScript", "features": [ "Server发现", "版本验证", "安全审计" ], "sources": [ { "source_type": "web", "source_credibility": "A", "collected_at": "2026-05-30T13:38:30.808Z" } ] }, { "id": "MCP-mongodb-v2", "name": "MongoDB MCP Server v2", "full_name": "@mongodb/mcp-server", "category": "数据库", "description": "MongoDB官方MCP Server升级版，支持Atlas集群管理", "protocol_version": "2025-11-25", "auth": "API Key", "hosts": [ "Claude Desktop", "Cursor" ], "security": "API Key + TLS", "stars": 1500, "maintainer": "MongoDB", "status": "active", "language": "TypeScript", "features": [ "CRUD操作", "聚合管道", "Atlas管理", "索引优化" ], "sources": [ { "source_type": "web", "source_credibility": "A", "collected_at": "2026-05-30T13:38:30.808Z" } ] }, { "id": "MCP-firecrawl-v2", "name": "Firecrawl MCP Server v2", "full_name": "@firecrawl/mcp-server", "category": "网页抓取", "description": "Firecrawl网页抓取MCP Server升级版，支持批量爬取和结构化提取", "protocol_version": "2025-11-25", "auth": "API Key", "hosts": [ "Claude Desktop", "Cursor" ], "security": "API Key", "stars": 4500, "maintainer": "Firecrawl", "status": "active", "language": "TypeScript", "features": [ "网页抓取", "批量爬取", "结构化提取", "JavaScript渲染" ], "sources": [ { "source_type": "web", "source_credibility": "A", "collected_at": "2026-05-30T13:38:30.808Z" } ] }, { "id": "MCP-redis-v2", "name": "Redis MCP Server v2", "full_name": "@redis/mcp-server", "category": "数据库", "description": "Redis MCP Server升级版，支持Redis Stack全功能", "protocol_version": "2025-11-25", "auth": "local", "hosts": [ "Claude Desktop", "Cursor" ], "security": "ACL", "stars": 900, "maintainer": "Redis", "status": "active", "language": "TypeScript", "features": [ "键值操作", "发布订阅", "流处理", "RediSearch" ], "sources": [ { "source_type": "web", "source_credibility": "A", "collected_at": "2026-05-30T13:38:30.808Z" } ] }, { "id": "MCP-coinbase", "name": "Coinbase MCP Server", "full_name": "coinbase-mcp", "category": "加密货币", "description": "Coinbase加密货币交易所MCP Server", "protocol_version": "2025-11-25", "auth": "API Key", "hosts": [ "Claude Desktop" ], "security": "API Key + HMAC", "stars": 600, "maintainer": "Coinbase", "status": "active", "language": "TypeScript", "features": [ "交易操作", "钱包查询", "市场数据" ], "sources": [ { "source_type": "web", "source_credibility": "A", "collected_at": "2026-05-30T13:38:30.808Z" } ] }, { "id": "MCP-stripe-v2", "name": "Stripe MCP Server v2", "full_name": "@stripe/mcp-server-v2", "category": "支付", "description": "Stripe支付MCP Server升级版，支持x402协议", "protocol_version": "2025-11-25", "auth": "API Key", "hosts": [ "Claude Desktop", "Cursor" ], "security": "API Key", "stars": 2500, "maintainer": "Stripe", "status": "active", "language": "TypeScript", "features": [ "支付处理", "订阅管理", "x402协议", "发票管理" ], "sources": [ { "source_type": "web", "source_credibility": "A", "collected_at": "2026-05-30T13:38:30.808Z" } ] }, { "id": "MCP-terraform-v2", "name": "Terraform MCP Server v2", "full_name": "@hashicorp/terraform-mcp", "category": "基础设施", "description": "HashiCorp Terraform MCP Server升级版，支持AI辅助IaC", "protocol_version": "2025-11-25", "auth": "API Key", "hosts": [ "Claude Desktop", "Cursor" ], "security": "API Key", "stars": 1200, "maintainer": "HashiCorp", "status": "active", "language": "Go", "features": [ "基础设施管理", "AI辅助配置", "Plan/Apply", "状态查询" ], "sources": [ { "source_type": "web", "source_credibility": "A", "collected_at": "2026-05-30T13:38:30.808Z" } ] }, { "id": "MCP-salesforce", "name": "Salesforce MCP Server", "full_name": "salesforce-mcp", "category": "CRM", "description": "Salesforce CRM平台MCP Server", "protocol_version": "2025-11-25", "auth": "OAuth", "hosts": [ "Claude Desktop", "Cursor" ], "security": "OAuth 2.0", "stars": 700, "maintainer": "社区/Salesforce", "status": "active", "language": "TypeScript", "features": [ "CRM操作", "SOQL查询", "对象管理", "流程自动化" ], "sources": [ { "source_type": "web", "source_credibility": "B", "collected_at": "2026-05-30T13:38:30.808Z" } ] }, { "id": "MCP-heroku", "name": "Heroku MCP Server", "full_name": "heroku-mcp", "category": "云平台", "description": "Heroku云平台MCP Server", "protocol_version": "2025-11-25", "auth": "API Key", "hosts": [ "Claude Desktop" ], "security": "API Key", "stars": 300, "maintainer": "社区", "status": "active", "language": "TypeScript", "features": [ "应用管理", "部署操作", "日志查看", "扩缩容" ], "sources": [ { "source_type": "web", "source_credibility": "B", "collected_at": "2026-05-30T13:38:30.808Z" } ] }, { "id": "MCP-supabase-v2", "name": "Supabase MCP Server v2", "full_name": "@supabase/mcp-server-v2", "category": "数据库", "description": "Supabase MCP Server升级版，支持Edge Functions和Storage", "protocol_version": "2025-11-25", "auth": "API Key", "hosts": [ "Claude Desktop", "Cursor" ], "security": "API Key", "stars": 2000, "maintainer": "Supabase", "status": "active", "language": "TypeScript", "features": [ "数据库操作", "Edge Functions", "Storage管理", "Auth管理" ], "sources": [ { "source_type": "web", "source_credibility": "A", "collected_at": "2026-05-30T13:38:30.808Z" } ] }, { "id": "MCP-cloudflare-v2", "name": "Cloudflare MCP Server v2", "full_name": "@cloudflare/mcp-server-v2", "category": "云平台", "description": "Cloudflare MCP Server升级版，支持Workers和Pages部署", "protocol_version": "2025-11-25", "auth": "API Key", "hosts": [ "Claude Desktop", "Cursor" ], "security": "API Key", "stars": 1800, "maintainer": "Cloudflare", "status": "active", "language": "TypeScript", "features": [ "Workers部署", "DNS管理", "CDN配置", "Pages发布" ], "sources": [ { "source_type": "web", "source_credibility": "A", "collected_at": "2026-05-30T13:38:30.808Z" } ] }, { "id": "MCP-notion-v2", "name": "Notion MCP Server v2", "full_name": "@notion/mcp-server-v2", "category": "知识管理", "description": "Notion MCP Server升级版，支持数据库操作和AI增强", "protocol_version": "2025-11-25", "auth": "OAuth", "hosts": [ "Claude Desktop", "Cursor" ], "security": "OAuth 2.0", "stars": 3500, "maintainer": "Notion", "status": "active", "language": "TypeScript", "features": [ "页面读写", "数据库操作", "搜索", "AI增强" ], "sources": [ { "source_type": "web", "source_credibility": "A", "collected_at": "2026-05-30T13:38:30.808Z" } ] }, { "id": "MCP-slack-v2", "name": "Slack MCP Server v2", "full_name": "@slack/mcp-server-v2", "category": "通信", "description": "Slack MCP Server升级版，支持频道管理和消息搜索", "protocol_version": "2025-11-25", "auth": "OAuth", "hosts": [ "Claude Desktop", "Cursor" ], "security": "OAuth 2.0", "stars": 2000, "maintainer": "Slack", "status": "active", "language": "TypeScript", "features": [ "消息发送", "频道管理", "文件上传", "搜索" ], "sources": [ { "source_type": "web", "source_credibility": "A", "collected_at": "2026-05-30T13:38:30.808Z" } ] }, { "id": "MCP-aws-v2", "name": "AWS MCP Server v2", "full_name": "@aws/mcp-server-v2", "category": "云平台", "description": "AWS MCP Server升级版，支持多服务集成", "protocol_version": "2025-11-25", "auth": "IAM", "hosts": [ "Claude Desktop", "Cursor" ], "security": "IAM Role", "stars": 3000, "maintainer": "AWS", "status": "active", "language": "TypeScript", "features": [ "S3操作", "Lambda管理", "EC2管理", "CloudWatch" ], "sources": [ { "source_type": "web", "source_credibility": "A", "collected_at": "2026-05-30T13:38:30.808Z" } ] }, { "id": "MCP-playwright-v2", "name": "Playwright MCP Server v2", "full_name": "@playwright/mcp-server-v2", "category": "浏览器自动化", "description": "Playwright MCP Server升级版，支持AI驱动的浏览器操作", "protocol_version": "2025-11-25", "auth": "local", "hosts": [ "Claude Desktop", "Cursor", "VS Code" ], "security": "本地沙箱", "stars": 5000, "maintainer": "Microsoft", "status": "active", "language": "TypeScript", "features": [ "浏览器自动化", "截图", "表单填写", "AI驱动操作" ], "sources": [ { "source_type": "web", "source_credibility": "A", "collected_at": "2026-05-30T13:38:30.808Z" } ] }, { "id": "MCP-atlassian-rovo", "name": "Atlassian Rovo MCP Server", "full_name": "@atlassian/rovo-mcp-server", "category": "项目管理", "description": "Atlassian官方云托管MCP Server，连接Jira、Confluence、Compass，单一安全端点", "protocol_version": "2025-11-25", "auth": "OAuth 2.0", "hosts": [ "Claude Desktop", "Cursor", "VS Code" ], "security": "OAuth 2.0 / 云端托管", "stars": 2200, "maintainer": "Atlassian", "status": "active", "language": "TypeScript", "features": [ "Jira Issue管理", "Confluence文档搜索", "Compass服务发现", "云托管零配置" ], "sources": [ { "source_type": "web", "source_credibility": "A", "collected_at": "2026-05-31T13:41:37.338Z" } ] }, { "id": "MCP-servicenow", "name": "ServiceNow MCP Server", "full_name": "servicenow-mcp-server", "category": "ITSM", "description": "ServiceNow IT服务管理MCP Server，支持Incident和Change管理", "protocol_version": "2025-11-25", "auth": "OAuth 2.0", "hosts": [ "Claude Desktop" ], "security": "OAuth 2.0", "stars": 600, "maintainer": "ServiceNow", "status": "active", "language": "TypeScript", "features": [ "Incident管理", "Change管理", "CMDB查询", "知识库搜索" ], "sources": [ { "source_type": "web", "source_credibility": "B", "collected_at": "2026-05-31T13:41:37.338Z" } ] }, { "id": "MCP-grafana", "name": "Grafana MCP Server", "full_name": "grafana-mcp-server", "category": "监控", "description": "Grafana可观测性平台MCP Server，支持Dashboard和告警管理", "protocol_version": "2025-11-25", "auth": "API Key", "hosts": [ "Claude Desktop" ], "security": "API Key", "stars": 700, "maintainer": "Grafana Labs", "status": "active", "language": "TypeScript", "features": [ "Dashboard管理", "数据源查询", "告警规则", "注解操作" ], "sources": [ { "source_type": "web", "source_credibility": "B", "collected_at": "2026-05-31T13:41:37.338Z" } ] }, { "id": "MCP-portkey-agentforce", "name": "Portkey AgentForce MCP", "full_name": "@portkey/agentforce-mcp", "category": "AI网关", "description": "Portkey提供的AgentForce连接方案，通过MCP连接Salesforce等企业系统", "protocol_version": "2025-11-25", "auth": "API Key", "hosts": [ "Claude Desktop", "Cursor" ], "security": "API Key + 网关", "stars": 500, "maintainer": "Portkey", "status": "active", "language": "TypeScript", "features": [ "企业系统连接", "AI网关", "多模型路由", "可观测性" ], "sources": [ { "source_type": "web", "source_credibility": "B", "collected_at": "2026-05-31T13:41:37.338Z" } ] }, { "id": "MCP-ahrefs-v2", "name": "Ahrefs MCP Server v2", "full_name": "@ahrefs/mcp-server", "category": "SEO", "description": "Ahrefs SEO分析工具MCP Server升级版，支持更丰富的SEO数据查询", "protocol_version": "2025-11-25", "auth": "API Key", "hosts": [ "Claude Desktop", "Cursor" ], "security": "API Key", "stars": 800, "maintainer": "Ahrefs", "status": "active", "language": "TypeScript", "features": [ "关键词研究", "反向链接分析", "SEO审计", "竞品分析", "排名追踪" ], "sources": [ { "source_type": "web", "source_credibility": "A", "collected_at": "2026-05-31T13:41:37.338Z" } ] }, { "id": "MCP-shopify-v2", "name": "Shopify MCP Server v2", "full_name": "@shopify/mcp-server", "category": "电商", "description": "Shopify官方MCP Server，支持产品/订单/库存/客户管理", "protocol_version": "2025-11-25", "auth": "OAuth", "hosts": [ "Claude Desktop", "Cursor" ], "security": "OAuth 2.0", "stars": 1500, "maintainer": "Shopify", "status": "active", "language": "TypeScript", "features": [ "产品管理", "订单处理", "库存同步", "客户分析", "折扣管理" ], "sources": [ { "source_type": "web", "source_credibility": "A", "collected_at": "2026-05-31T13:41:37.338Z" } ] }, { "id": "MCP-hubspot-v2", "name": "HubSpot MCP Server v2", "full_name": "@hubspot/mcp-server", "category": "CRM", "description": "HubSpot官方MCP Server，支持CRM全功能操作", "protocol_version": "2025-11-25", "auth": "OAuth", "hosts": [ "Claude Desktop", "Cursor" ], "security": "OAuth 2.0", "stars": 1000, "maintainer": "HubSpot", "status": "active", "language": "TypeScript", "features": [ "CRM操作", "营销自动化", "销售管道", "客户服务", "内容管理" ], "sources": [ { "source_type": "web", "source_credibility": "A", "collected_at": "2026-05-31T13:41:37.338Z" } ] }, { "id": "MCP-vercel-ai-gateway", "name": "Vercel AI Gateway MCP", "full_name": "@vercel/ai-gateway-mcp", "category": "AI网关", "description": "Vercel AI Gateway MCP Server，统一多模型AI访问", "protocol_version": "2025-11-25", "auth": "API Token", "hosts": [ "Claude Desktop", "Cursor", "VS Code" ], "security": "Token权限", "stars": 1200, "maintainer": "Vercel", "status": "active", "language": "TypeScript", "features": [ "多模型路由", "流式响应", "缓存优化", "Fallback策略" ], "sources": [ { "source_type": "web", "source_credibility": "A", "collected_at": "2026-05-31T13:41:37.338Z" } ] }, { "id": "MCP-scrapling", "name": "Scrapling MCP Server", "category": "mcp_servers", "description": "Adaptive web scraping framework with native MCP server integration for AI agents", "github": "https://github.com/D4Vinci/Scrapling", "pypi": "scrapling", "version": "0.4.8", "language": "Python", "key_features": [ "Adaptive element tracking - survives website redesigns", "Native anti-bot bypass (Cloudflare Turnstile etc.)", "Three-tier Fetcher system (Fetcher/StealthyFetcher/DynamicFetcher)", "Scrapy-like Spider framework with concurrent crawling", "MCP Server for conversational scraping with AI agents", "Auto-resuming spiders with pause/recovery", "Real-time statistics and streaming" ], "fetcher_types": { "Fetcher": "Basic HTTP requests", "StealthyFetcher": "Stealth mode bypassing anti-bot detection", "DynamicFetcher": "Handles JS-rendered pages" }, "use_cases": [ "Data collection for knowledge bases", "Anti-bot protected site scraping", "Large-scale concurrent crawling", "AI agent web interaction", "Website change monitoring" ], "adaptive_feature": "Elements tracked by smart structure analysis, not fragile CSS selectors - survives website redesigns automatically", "anti_bot": [ "Cloudflare Turnstile", "DataDome", "PerimeterX", "Akamai Bot Manager" ], "installation": "pip install scrapling", "mcp_setup": "Built-in MCP server for Claude/AI agent integration", "stars": "5k+", "license": "BSD-3", "sources": [ "web_search", "github" ], "confidence": "high", "first_seen": "2026-05-31T22:02:40.222Z", "last_updated": "2026-05-31T22:02:40.222Z" }, { "id": "MCP-atlassian-remote", "name": "Atlassian Remote MCP Server", "full_name": "@atlassian/mcp-server", "category": "项目管理", "description": "Atlassian官方远程MCP Server，覆盖Jira、Confluence、Compass，2026年2月GA", "protocol_version": "2025-11-25", "auth": "OAuth", "hosts": [ "Claude Desktop", "Cursor", "VS Code" ], "security": "OAuth 2.0", "stars": 2100, "maintainer": "Atlassian", "status": "active", "language": "TypeScript", "features": [ "Jira操作", "Confluence文档", "Compass组件", "远程部署" ], "sources": [ { "source_type": "web", "source_credibility": "A", "collected_at": "2026-06-01T13:48:20.198Z" } ] }, { "id": "MCP-codex", "name": "OpenAI Codex MCP Server", "full_name": "@openai/codex-mcp", "category": "代码", "description": "OpenAI Codex代码生成和执行MCP Server", "protocol_version": "2025-11-25", "auth": "API Key", "hosts": [ "Claude Desktop", "Cursor" ], "security": "API Key", "stars": 1800, "maintainer": "OpenAI", "status": "active", "language": "TypeScript", "features": [ "代码生成", "代码执行", "代码审查", "多语言支持" ], "sources": [ { "source_type": "web", "source_credibility": "A", "collected_at": "2026-06-01T13:48:20.198Z" } ] }, { "id": "MCP-truto", "name": "Truto MCP Server", "full_name": "truto-mcp-server", "category": "API", "description": "统一API集成平台MCP Server，支持100+ SaaS应用", "protocol_version": "2025-11-25", "auth": "API Key", "hosts": [ "Claude Desktop", "Cursor" ], "security": "API Key", "stars": 600, "maintainer": "Truto", "status": "active", "language": "TypeScript", "features": [ "100+ SaaS集成", "统一API", "数据同步", "CRUD操作" ], "sources": [ { "source_type": "web", "source_credibility": "B", "collected_at": "2026-06-01T13:48:20.198Z" } ] }, { "id": "MCP-skyvia-v2", "name": "Skyvia MCP Server v2", "full_name": "skyvia-mcp-v2", "category": "数据集成", "description": "Skyvia数据集成平台MCP Server升级版，支持实时数据管道", "protocol_version": "2025-11-25", "auth": "API Key", "hosts": [ "Claude Desktop", "Cursor" ], "security": "OAuth 2.0", "stars": 900, "maintainer": "Skyvia", "status": "active", "language": "TypeScript", "features": [ "实时数据管道", "ETL/ELT", "云数据库同步", "数据备份" ], "sources": [ { "source_type": "web", "source_credibility": "B", "collected_at": "2026-06-01T13:48:20.198Z" } ] }, { "id": "MCP-prefect", "name": "Prefect MCP Server", "full_name": "prefect-mcp-server", "category": "工作流", "description": "Prefect工作流编排平台MCP Server，AI Agent工作流管理", "protocol_version": "2025-11-25", "auth": "API Key", "hosts": [ "Claude Desktop", "Cursor" ], "security": "API Key", "stars": 700, "maintainer": "Prefect", "status": "active", "language": "Python", "features": [ "工作流编排", "任务调度", "监控告警", "数据管道" ], "sources": [ { "source_type": "web", "source_credibility": "B", "collected_at": "2026-06-01T13:48:20.198Z" } ] }, { "id": "MCP-titanapps", "name": "Titan Apps Jira MCP", "full_name": "titanapps-jira-mcp", "category": "项目管理", "description": "云端托管的Jira MCP Server，支持Jira/Confluence/Compass", "protocol_version": "2025-11-25", "auth": "OAuth", "hosts": [ "Claude Desktop", "Cursor" ], "security": "云端OAuth", "stars": 400, "maintainer": "Titan Apps", "status": "active", "language": "TypeScript", "features": [ "云端部署", "Jira集成", "Confluence集成", "无需本地安装" ], "sources": [ { "source_type": "web", "source_credibility": "B", "collected_at": "2026-06-01T13:48:20.198Z" } ] }, { "id": "MCP-instatunnel", "name": "Instatunnel MCP Server", "full_name": "instatunnel-mcp", "category": "安全", "description": "MCP Server安全隧道，保护远程MCP连接", "protocol_version": "2025-11-25", "auth": "API Key", "hosts": [ "Claude Desktop" ], "security": "端到端加密", "stars": 300, "maintainer": "Instatunnel", "status": "active", "language": "TypeScript", "features": [ "安全隧道", "远程访问", "加密传输", "审计日志" ], "sources": [ { "source_type": "web", "source_credibility": "B", "collected_at": "2026-06-01T13:48:20.198Z" } ] }, { "id": "MCP-mcpmarket", "name": "MCPMarket Registry", "full_name": "mcpmarket-registry", "category": "发现", "description": "MCP Server市场与注册中心，每日更新排名", "protocol_version": "2025-11-25", "auth": "none", "hosts": [ "Web" ], "security": "公开", "stars": 500, "maintainer": "MCPMarket", "status": "active", "language": "Web", "features": [ "Server发现", "排名追踪", "社区评价", "每日更新" ], "sources": [ { "source_type": "web", "source_credibility": "B", "collected_at": "2026-06-01T13:48:20.198Z" } ] }, { "id": "MCP-totalum", "name": "Totalum MCP Server", "full_name": "totalum-mcp", "category": "CRM", "description": "Totalum CRM/ERP平台MCP Server", "protocol_version": "2025-11-25", "auth": "API Key", "hosts": [ "Claude Desktop", "Cursor" ], "security": "API Key", "stars": 200, "maintainer": "Totalum", "status": "active", "language": "TypeScript", "features": [ "CRM操作", "ERP集成", "数据管理", "报表生成" ], "sources": [ { "source_type": "web", "source_credibility": "C", "collected_at": "2026-06-01T13:48:20.198Z" } ] }, { "id": "MCP-tokenmix", "name": "TokenMix MCP Directory", "full_name": "tokenmix-mcp-directory", "category": "发现", "description": "70+生产级MCP Server目录，2026完整指南", "protocol_version": "2025-11-25", "auth": "none", "hosts": [ "Web" ], "security": "公开", "stars": 300, "maintainer": "TokenMix", "status": "active", "language": "Web", "features": [ "Server目录", "生产级筛选", "使用指南", "分类浏览" ], "sources": [ { "source_type": "web", "source_credibility": "B", "collected_at": "2026-06-01T13:48:20.198Z" } ] }, { "id": "MCP-cf-d1", "name": "Cloudflare D1 Remote MCP Server", "type": "mcp_server", "category": "database", "description": "Cloudflare D1 SQLite数据库远程MCP Server，2026年4月发布，支持AI Agent直接查询D1数据库", "vendor": "Cloudflare", "protocol": "MCP", "deployment": "remote", "release_date": "2026-04", "source": "web_search_2026" }, { "id": "MCP-cf-r2", "name": "Cloudflare R2 Remote MCP Server", "type": "mcp_server", "category": "storage", "description": "Cloudflare R2对象存储远程MCP Server，2026年4月发布，支持AI Agent读写R2存储桶", "vendor": "Cloudflare", "protocol": "MCP", "deployment": "remote", "release_date": "2026-04", "source": "web_search_2026" }, { "id": "MCP-cf-workers-logs", "name": "Cloudflare Workers Logs MCP Server", "type": "mcp_server", "category": "monitoring", "description": "Cloudflare Workers日志远程MCP Server，2026年4月发布，支持AI Agent查询和分析Workers运行日志", "vendor": "Cloudflare", "protocol": "MCP", "deployment": "remote", "release_date": "2026-04", "source": "web_search_2026" }, { "id": "MCP-cf-containers", "name": "Cloudflare Containers MCP Server", "type": "mcp_server", "category": "compute", "description": "Cloudflare容器服务远程MCP Server，2026年4月发布，支持AI Agent管理容器化工作负载", "vendor": "Cloudflare", "protocol": "MCP", "deployment": "remote", "release_date": "2026-04", "source": "web_search_2026" }, { "id": "MCP-cf-kv", "name": "Cloudflare KV Remote MCP Server", "type": "mcp_server", "category": "database", "description": "Cloudflare KV键值存储远程MCP Server，支持AI Agent读写全球分布式KV存储", "vendor": "Cloudflare", "protocol": "MCP", "deployment": "remote", "release_date": "2026-04", "source": "web_search_2026" }, { "id": "MCP-cf-queues", "name": "Cloudflare Queues MCP Server", "type": "mcp_server", "category": "messaging", "description": "Cloudflare消息队列远程MCP Server，支持AI Agent发送和消费队列消息", "vendor": "Cloudflare", "protocol": "MCP", "deployment": "remote", "release_date": "2026-04", "source": "web_search_2026" }, { "id": "MCP-cf-durable-objects", "name": "Cloudflare Durable Objects MCP Server", "type": "mcp_server", "category": "compute", "description": "Cloudflare Durable Objects远程MCP Server，支持AI Agent管理有状态计算对象", "vendor": "Cloudflare", "protocol": "MCP", "deployment": "remote", "release_date": "2026-04", "source": "web_search_2026" }, { "id": "MCP-cf-workers-ai", "name": "Cloudflare Workers AI MCP Server", "type": "mcp_server", "category": "ai_inference", "description": "Cloudflare Workers AI推理远程MCP Server，支持AI Agent调用边缘AI模型推理", "vendor": "Cloudflare", "protocol": "MCP", "deployment": "remote", "release_date": "2026-04", "source": "web_search_2026" }, { "id": "MCP-cf-pages", "name": "Cloudflare Pages MCP Server", "type": "mcp_server", "category": "deployment", "description": "Cloudflare Pages部署远程MCP Server，支持AI Agent管理静态站点和全栈应用部署", "vendor": "Cloudflare", "protocol": "MCP", "deployment": "remote", "release_date": "2026-04", "source": "web_search_2026" }, { "id": "MCP-cf-images", "name": "Cloudflare Images MCP Server", "type": "mcp_server", "category": "media", "description": "Cloudflare图片服务远程MCP Server，支持AI Agent上传、转换和管理图片", "vendor": "Cloudflare", "protocol": "MCP", "deployment": "remote", "release_date": "2026-04", "source": "web_search_2026" }, { "id": "MCP-cf-stream", "name": "Cloudflare Stream MCP Server", "type": "mcp_server", "category": "media", "description": "Cloudflare视频流远程MCP Server，支持AI Agent上传和管理视频内容", "vendor": "Cloudflare", "protocol": "MCP", "deployment": "remote", "release_date": "2026-04", "source": "web_search_2026" }, { "id": "MCP-cf-email-workers", "name": "Cloudflare Email Workers MCP Server", "type": "mcp_server", "category": "email", "description": "Cloudflare邮件Workers远程MCP Server，支持AI Agent处理和发送电子邮件", "vendor": "Cloudflare", "protocol": "MCP", "deployment": "remote", "release_date": "2026-04", "source": "web_search_2026" }, { "id": "MCP-cf-turnstile", "name": "Cloudflare Turnstile MCP Server", "type": "mcp_server", "category": "security", "description": "Cloudflare Turnstile验证远程MCP Server，支持AI Agent管理CAPTCHA验证和机器人防护", "vendor": "Cloudflare", "protocol": "MCP", "deployment": "remote", "release_date": "2026-04", "source": "web_search_2026" }, { "id": "MCP-exa-search", "name": "Exa Real-Time Web Search MCP Server", "type": "mcp_server", "category": "search", "description": "Exa高级实时网络搜索MCP Server，支持语义搜索和实时信息检索，2026年Vibe Coding热门工具", "vendor": "Exa", "protocol": "MCP", "deployment": "remote", "release_date": "2025", "source": "web_search_2026" }, { "id": "MCP-context7-docs", "name": "Context7 Documentation MCP Server", "type": "mcp_server", "category": "documentation", "description": "Context7最新代码文档MCP Server，支持AI Agent获取最新的库和框架文档，2026年Vibe Coding热门工具", "vendor": "Context7", "protocol": "MCP", "deployment": "remote", "release_date": "2025", "source": "web_search_2026" }, { "id": "MCP-firecrawl-scraping", "name": "Firecrawl Web Scraping MCP Server", "type": "mcp_server", "category": "web_scraping", "description": "Firecrawl网页抓取MCP Server，支持AI Agent抓取和解析网页内容，2026年开发者Top 10 MCP Server", "vendor": "Firecrawl", "protocol": "MCP", "deployment": "remote", "release_date": "2025", "source": "web_search_2026" }, { "id": "MCP-browserbase", "name": "Browserbase Browser Automation MCP Server", "type": "mcp_server", "category": "browser_automation", "description": "Browserbase浏览器自动化MCP Server，支持AI Agent在云端浏览器中执行操作，2026年开发者Top 10", "vendor": "Browserbase", "protocol": "MCP", "deployment": "remote", "release_date": "2025", "source": "web_search_2026" }, { "id": "mcp-speakeasy", "name": "Speakeasy MCP Server", "type": "api_tooling", "description": "Speakeasy提供的MCP Server，用于API开发和SDK生成，支持MCP协议规范发布管理", "company": "Speakeasy", "category": "developer_tools", "key_features": [ "API SDK生成", "MCP规范发布管理", "多语言SDK支持", "API测试" ], "status": "production", "year": 2026, "sources": [ { "source_type": "web", "source_credibility": "B", "collected_at": "2026-06-03T13:52:48.985531Z" } ] }, { "id": "mcp-pgedge", "name": "pgEdge MCP Server", "type": "database", "description": "pgEdge分布式PostgreSQL的MCP Server，支持多区域数据库管理和低延迟数据访问", "company": "pgEdge", "category": "database", "key_features": [ "分布式PostgreSQL", "多区域复制", "低延迟访问", "SQL操作" ], "status": "production", "year": 2026, "sources": [ { "source_type": "web", "source_credibility": "B", "collected_at": "2026-06-03T13:52:48.985531Z" } ] }, { "id": "mcp-scalekit", "name": "Scalekit MCP Server", "type": "auth_benchmark", "description": "Scalekit认证和基准测试MCP Server，提供MCP Server性能基准测试数据", "company": "Scalekit", "category": "developer_tools", "key_features": [ "MCP性能基准", "认证集成", "SAML/SSO", "多租户" ], "status": "production", "year": 2026, "sources": [ { "source_type": "web", "source_credibility": "B", "collected_at": "2026-06-03T13:52:48.985531Z" } ] }, { "id": "mcp-composio", "name": "Composio MCP Server", "type": "tool_integration", "description": "Composio工具集成平台MCP Server，支持250+工具和应用集成，Agent工具调用中间件", "company": "Composio", "category": "integration", "key_features": [ "250+工具集成", "Agent工具调用", "OAuth管理", "多平台支持" ], "status": "production", "year": 2026, "sources": [ { "source_type": "web", "source_credibility": "B", "collected_at": "2026-06-03T13:52:48.985531Z" } ] }, { "id": "mcp-smithery", "name": "Smithery MCP Registry", "type": "registry", "description": "Smithery MCP Server市场，MCP Server发现和部署平台，提供服务器目录和一键部署", "company": "Smithery", "category": "registry", "key_features": [ "MCP Server市场", "一键部署", "服务器发现", "版本管理" ], "status": "production", "year": 2026, "sources": [ { "source_type": "web", "source_credibility": "B", "collected_at": "2026-06-03T13:52:48.985531Z" } ] }, { "id": "mcp-cursor", "name": "Cursor MCP Server", "type": "ide_integration", "description": "Cursor IDE的MCP Server集成，支持AI编程助手的工具调用和代码上下文管理", "company": "Cursor/Anysphere", "category": "developer_tools", "key_features": [ "IDE集成", "代码上下文", "AI编程助手", "工具调用" ], "status": "production", "year": 2026, "sources": [ { "source_type": "web", "source_credibility": "A", "collected_at": "2026-06-03T13:52:48.985531Z" } ] }, { "id": "mcp-claude-desktop", "name": "Claude Desktop MCP Host", "type": "desktop_host", "description": "Claude Desktop作为MCP Host，支持连接多个MCP Server，是MCP生态最广泛使用的客户端", "company": "Anthropic", "category": "host", "key_features": [ "MCP Host客户端", "多Server连接", "桌面集成", "工具发现" ], "status": "production", "year": 2026, "sources": [ { "source_type": "web", "source_credibility": "A", "collected_at": "2026-06-03T13:52:48.985531Z" } ] }, { "id": "mcp-replit", "name": "Replit MCP Server", "type": "development", "description": "Replit开发环境的MCP Server，支持云端开发环境的工具调用和代码执行", "company": "Replit", "category": "developer_tools", "key_features": [ "云端开发", "代码执行", "环境管理", "协作开发" ], "status": "production", "year": 2026, "sources": [ { "source_type": "web", "source_credibility": "B", "collected_at": "2026-06-03T13:52:48.985531Z" } ] }, { "id": "mcp-stainless", "name": "Stainless MCP Server", "type": "api_tooling", "description": "Stainless API标准MCP Server，支持API规范生成和SDK自动构建", "company": "Stainless", "category": "developer_tools", "key_features": [ "API规范生成", "SDK自动构建", "OpenAPI支持", "多语言" ], "status": "production", "year": 2026, "sources": [ { "source_type": "web", "source_credibility": "B", "collected_at": "2026-06-03T13:52:48.985531Z" } ] }, { "id": "mcp-tyk-security", "name": "Tyk MCP Security Gateway", "type": "security", "description": "Tyk企业MCP安全网关，提供MCP Server安全策略执行、认证和监控", "company": "Tyk", "category": "security", "key_features": [ "安全策略执行", "认证管理", "性能监控", "企业级" ], "status": "production", "year": 2026, "sources": [ { "source_type": "web", "source_credibility": "B", "collected_at": "2026-06-03T13:52:48.985531Z" } ] }, { "id": "MCP-stackgen", "name": "StackGen MCP Server", "full_name": "stackgen-mcp", "category": "部署", "description": "StackGen部署自动化MCP Server，减少部署时间从天到分钟", "protocol_version": "2025-11-25", "auth": "API Key", "hosts": [ "Claude Desktop", "Cursor" ], "security": "API Key", "stars": 400, "maintainer": "StackGen", "status": "active", "language": "TypeScript", "features": [ "部署自动化", "基础设施生成", "分钟级部署" ], "sources": [ { "source_type": "web", "source_credibility": "B", "collected_at": "2026-06-04T14:03:15.064Z" } ] }, { "id": "MCP-mcpmanager", "name": "MCP Manager Server", "full_name": "mcpmanager-mcp", "category": "管理", "description": "MCP Server管理工具，支持发现、安装和配置MCP Server", "protocol_version": "2025-11-25", "auth": "local", "hosts": [ "Claude Desktop" ], "security": "本地沙箱", "stars": 800, "maintainer": "社区", "status": "active", "language": "TypeScript", "features": [ "Server发现", "安装管理", "配置同步" ], "sources": [ { "source_type": "web", "source_credibility": "B", "collected_at": "2026-06-04T14:03:15.064Z" } ] }, { "id": "MCP-workos", "name": "WorkOS MCP Server", "full_name": "workos-mcp", "category": "认证", "description": "WorkOS身份认证MCP Server，支持SSO、目录同步和身份管理", "protocol_version": "2025-11-25", "auth": "API Key", "hosts": [ "Claude Desktop", "Cursor" ], "security": "API Key", "stars": 500, "maintainer": "WorkOS", "status": "active", "language": "TypeScript", "features": [ "SSO认证", "目录同步", "身份管理" ], "sources": [ { "source_type": "web", "source_credibility": "B", "collected_at": "2026-06-04T14:03:15.064Z" } ] }, { "id": "MCP-shareuhack", "name": "Shareuhack Dev MCP", "full_name": "shareuhack-dev-mcp", "category": "开发", "description": "开发者工具MCP Server集合，集成多种开发辅助工具", "protocol_version": "2025-11-25", "auth": "none", "hosts": [ "Claude Desktop", "VS Code" ], "security": "本地沙箱", "stars": 300, "maintainer": "社区", "status": "active", "language": "TypeScript", "features": [ "代码辅助", "开发工具", "调试支持" ], "sources": [ { "source_type": "web", "source_credibility": "B", "collected_at": "2026-06-04T14:03:15.064Z" } ] }, { "id": "MCP-rasa", "name": "Rasa MCP Server", "full_name": "rasa-mcp", "category": "AI", "description": "Rasa对话AI平台MCP Server，支持企业级对话Agent构建", "protocol_version": "2025-11-25", "auth": "API Key", "hosts": [ "Claude Desktop" ], "security": "API Key", "stars": 700, "maintainer": "Rasa", "status": "active", "language": "Python", "features": [ "对话AI", "意图识别", "企业级NLU" ], "sources": [ { "source_type": "web", "source_credibility": "B", "collected_at": "2026-06-04T14:03:15.064Z" } ] }, { "id": "MCP-context7", "name": "Context7 MCP Server", "full_name": "context7-mcp", "category": "文档", "description": "Context7拉取最新版本特定文档和代码示例到LLM上下文，解决过时文档问题", "protocol_version": "2025-11-25", "auth": "API Key", "hosts": [ "Claude Desktop", "Cursor", "VS Code" ], "security": "API Key", "stars": 5000, "maintainer": "Context7", "status": "active", "language": "TypeScript", "features": [ "实时文档", "版本特定", "代码示例", "自动更新" ], "sources": [ { "source_type": "web", "source_credibility": "B", "collected_at": "2026-06-04T14:05:21.771Z" } ] }, { "id": "MCP-proxyman", "name": "Proxyman MCP Server", "full_name": "proxyman-mcp", "category": "调试", "description": "Proxyman网络调试MCP Server，支持HTTP/HTTPS流量检查和分析", "protocol_version": "2025-11-25", "auth": "local", "hosts": [ "Claude Desktop" ], "security": "本地沙箱", "stars": 1500, "maintainer": "Proxyman", "status": "active", "language": "TypeScript", "features": [ "网络调试", "流量分析", "HTTP检查", "HTTPS解密" ], "sources": [ { "source_type": "web", "source_credibility": "B", "collected_at": "2026-06-04T14:05:21.771Z" } ] }, { "id": "MCP-atlassian", "name": "Atlassian MCP Server", "full_name": "atlassian-mcp", "category": "项目管理", "description": "Atlassian Jira/Confluence集成MCP Server，支持项目管理和知识库操作", "protocol_version": "2025-11-25", "auth": "OAuth", "hosts": [ "Claude Desktop", "Cursor" ], "security": "OAuth 2.0", "stars": 2500, "maintainer": "Atlassian", "status": "active", "language": "TypeScript", "features": [ "Jira集成", "Confluence操作", "项目管理", "知识库" ], "sources": [ { "source_type": "web", "source_credibility": "B", "collected_at": "2026-06-04T14:05:21.771Z" } ] }, { "id": "MCP-tessl", "name": "Tessl MCP Server", "full_name": "tessl-mcp", "category": "开发", "description": "Tessl开发工具MCP Server，12种开发者生产力工具集合", "protocol_version": "2025-11-25", "auth": "API Key", "hosts": [ "Claude Desktop", "VS Code" ], "security": "API Key", "stars": 800, "maintainer": "Tessl", "status": "active", "language": "TypeScript", "features": [ "开发工具", "生产力", "多工具集成", "代码辅助" ], "sources": [ { "source_type": "web", "source_credibility": "B", "collected_at": "2026-06-04T14:05:21.771Z" } ] }, { "id": "MCP-playwright-mcp", "name": "Playwright MCP", "category": "browser-automation", "description": "Playwright MCP server ranked #1 globally in 2026, beating GitHub and Figma MCP servers. Enables browser automation via MCP protocol.", "features": [ "browser automation", "web scraping", "page interaction", "screenshot capture" ], "status": "active", "year": 2025, "sources": [ "https://mcpmanager.ai/blog/most-popular-mcp-servers", "https://www.reddit.com/r/mcp/comments/1s3fu45/top_50_most_popular_mcp_servers_in_2026" ] }, { "id": "MCP-registry", "name": "MCP Registry", "category": "registry", "description": "Official MCP Registry launched September 2025, growing to nearly 2,000 server entries within months. Central hub for discovering MCP servers.", "features": [ "server discovery", "registry", "metadata indexing", "community catalog" ], "status": "active", "year": 2025, "sources": [ "https://workos.com/blog/everything-your-team-needs-to-know-about-mcp-in-2026" ] }, { "id": "MCP-firecrawl", "name": "Firecrawl MCP Server", "category": "web-scraping", "description": "Web scraping MCP server by Firecrawl. Listed among top 10 MCP servers for developers in 2026.", "features": [ "web scraping", "data extraction", "crawling", "structured output" ], "status": "active", "year": 2025, "sources": [ "https://www.firecrawl.dev/blog/best-mcp-servers-for-developers" ] }, { "id": "MCP-knit", "name": "Knit MCP Server", "category": "integration", "description": "MCP integration server by GetKnit. Part of the 2026 MCP roadmap for AI integration.", "features": [ "AI integration", "tool connectivity", "data bridging" ], "status": "active", "year": 2026, "sources": [ "https://www.getknit.dev/blog/the-future-of-mcp-roadmap-enhancements-and-whats-next" ] }, { "id": "MCP-mcpmanager", "name": "MCP Manager", "category": "management", "description": "MCP server management platform tracking top 50 most popular MCP servers globally. Provides analytics and rankings.", "features": [ "server management", "analytics", "rankings", "popularity tracking" ], "status": "active", "year": 2026, "sources": [ "https://mcpmanager.ai/blog/most-popular-mcp-servers" ] }, { "id": "MCP-extensions-2026", "name": "MCP Extensions Framework", "full_name": "Model Context Protocol Extensions", "category": "协议扩展", "description": "MCP 2026-07-28 RC版本引入Extensions框架，允许社区扩展MCP核心协议而不破坏兼容性，支持自定义工具类型和资源格式", "protocol_version": "2026-07-28-RC", "auth": "varies", "hosts": [ "Claude Desktop", "Cursor", "VS Code", "所有MCP客户端" ], "security": "扩展沙箱", "stars": 0, "maintainer": "Anthropic + 社区", "status": "active", "language": "TypeScript/Python", "features": [ "自定义工具类型", "资源格式扩展", "协议兼容性保证", "社区扩展生态" ], "sources": [ { "source_type": "web", "source_credibility": "A", "article_url": "https://blog.modelcontextprotocol.io/posts/2026-07-28-release-candidate", "collected_at": "2026-06-16T22:31:38.000Z" } ] }, { "id": "MCP-stateless-core", "name": "MCP Stateless Protocol Core", "full_name": "MCP Stateless Core", "category": "协议核心", "description": "MCP 2026-07-28 RC版本引入无状态协议核心，解决有状态连接的可扩展性问题，支持HTTP无状态调用和大规模Agent部署", "protocol_version": "2026-07-28-RC", "auth": "varies", "hosts": [ "所有MCP客户端" ], "security": "无状态安全", "stars": 0, "maintainer": "Anthropic", "status": "active", "language": "TypeScript", "features": [ "无状态调用", "HTTP扩展", "大规模部署支持", "向后兼容" ], "sources": [ { "source_type": "web", "source_credibility": "A", "article_url": "https://blog.modelcontextprotocol.io/posts/2026-07-28-release-candidate", "collected_at": "2026-06-16T22:31:38.000Z" } ] }, { "id": "MCP-aishield-scanner", "name": "AIShield MCP Scanner", "full_name": "aishield-mcp", "category": "安全扫描", "description": "Agent原生AI工具安全扫描器。扫描MCP Server/GPT/Skill/Prompt的安全风险，4维评分(安全/隐私/质量/性能)，119条规则覆盖OWASP MCP Top 10", "protocol_version": "2025-03-26", "auth": "API Key (免费5次/天)", "hosts": [ "Claude Desktop", "Cursor", "VS Code", "Cline" ], "security": "4维评分+认证徽章", "stars": 0, "maintainer": "lm203688", "status": "active", "language": "JavaScript/Python", "features": [ "MCP安全扫描", "GPT安全扫描", "Prompt注入检测", "供应链分析", "凭证泄露检测", "认证徽章", "批量扫描" ], "install": "npx aishield-mcp", "github": "https://github.com/lm203688/aishield", "sources": [ { "source_type": "web", "source_credibility": "A", "article_url": "https://github.com/lm203688/aishield", "collected_at": "2026-06-18T09:19:49.000Z" } ] }, { "id": "MCP-aishield-guardrail", "name": "AIShield Guardrail MCP", "full_name": "aishield-guardrail", "category": "安全防护", "description": "杀手级产品：拦截Agent工具安装，自动触发安全扫描。放在MCP列表第一个，所有tool_install请求先扫描后放行", "protocol_version": "2025-03-26", "auth": "API Key (可选)", "hosts": [ "Claude Desktop", "Cursor", "VS Code", "Cline" ], "security": "主动拦截+自动阻断", "stars": 0, "maintainer": "lm203688", "status": "active", "language": "JavaScript", "features": [ "工具安装拦截", "自动安全扫描", "不安全工具阻断", "可配置阈值", "实时告警" ], "install": "npx aishield-guardrail", "github": "https://github.com/lm203688/aishield", "sources": [ { "source_type": "web", "source_credibility": "A", "article_url": "https://github.com/lm203688/aishield", "collected_at": "2026-06-18T09:19:49.000Z" } ] }, { "name": "GitHub MCP Server", "full_name": "Official GitHub MCP Server", "category": "Code Repository Integration", "description": "Official server providing GitHub repository integration for AI models, enabling code browsing, PR management, and issue tracking.", "protocol_version": "2025-04-public-preview", "auth": "OAuth (GitHub PAT)", "hosts": [ "GitHub", "Claude Desktop", "Cursor" ], "features": [ "Repository search", "PR review", "Issue management", "Code browsing" ], "sources": [ "[2]" ], "id": "MCP-164" }, { "name": "Anthropic MCP SDK", "full_name": "Model Context Protocol SDK (Anthropic)", "category": "Core SDK / Reference Implementation", "description": "Reference implementation and SDK for building MCP-compatible clients and servers, developed by Anthropic.", "protocol_version": "2025-03-26", "auth": "N/A", "hosts": [ "Anthropic Claude", "Open-source" ], "features": [ "Client library", "Server framework", "JSON-RPC transport", "Tool/resource/prompt definitions" ], "sources": [ "[7]", "[9]", "[10]" ], "id": "MCP-165" }, { "name": "WorkOS MCP Integration", "full_name": "WorkOS MCP Server", "category": "Authentication & User Management", "description": "Server integrating WorkOS authentication and user management capabilities into MCP-enabled AI applications.", "protocol_version": "2025-03-26", "auth": "WorkOS API Key + OAuth", "hosts": [ "WorkOS Platform" ], "features": [ "SSO/OIDC", "User provisioning", "Role-based access", "Directory sync" ], "sources": [ "[3]" ], "id": "MCP-166" }, { "name": "Linux Foundation MCP WG", "full_name": "Model Context Protocol Working Group (Linux Foundation)", "category": "Standardization / Governance Body", "description": "Multi-company open standard governance body under the Linux Foundation managing MCP specification evolution through Working Groups.", "protocol_version": "SEP-1302", "auth": "LF Membership", "hosts": [ "Linux Foundation", "modelcontextprotocol.io" ], "features": [ "Specification management", "Working groups", "Interest groups", "Release candidates" ], "sources": [ "[1]", "[4]" ], "id": "MCP-167" }, { "name": "Firecrawl MCP", "full_name": "Firecrawl Web Scraping MCP Server", "category": "Web Data Extraction", "description": "MCP server for web scraping, crawling, and extracting structured data from websites.", "protocol_version": "2025-03-26", "auth": "API Key", "hosts": [ "firecrawl.dev" ], "features": [ "Web scraping", "Browser automation", "Data extraction", "Site crawling" ], "sources": [ "[6]" ], "id": "MCP-168" }, { "name": "Speakeasy MCP", "full_name": "Speakeasy SDK Generation MCP Server", "category": "Developer Tooling", "description": "MCP server supporting API SDK generation, tool standardization, and authorization transport for developer workflows.", "protocol_version": "2025-03-26", "auth": "API Key / OAuth", "hosts": [ "speakeasy.com" ], "features": [ "SDK generation", "Tool standardization", "Authorization management", "Transport layer support" ], "sources": [ "[7]" ], "id": "MCP-169" }, { "name": "WorkOS MCP", "full_name": "WorkOS Identity & Auth MCP Server", "category": "Authentication & Identity", "description": "MCP server integrating AI models with external identity providers, authentication flows, and user management systems.", "protocol_version": "2025-03-26", "auth": "OAuth / SSO", "hosts": [ "workos.com" ], "features": [ "User authentication", "SSO integration", "Identity provisioning", "External tool connectivity" ], "sources": [ "[3]" ], "id": "MCP-170" }, { "name": "Anthropic Code Execution MCP", "full_name": "Anthropic Code Execution MCP Server", "category": "Code Execution & Agent Tools", "description": "MCP server enabling AI agents to execute code safely in sandboxed environments and interact with external systems programmatically.", "protocol_version": "2025-03-26", "auth": "Anthropic API Key", "hosts": [ "anthropic.com" ], "features": [ "Secure code execution", "Sandboxed runtime", "Agent tool integration", "Two-way data connection" ], "sources": [ "[10]", "[9]" ], "id": "MCP-171" }, { "id": "MCP-172", "name": "MCP 2026-07 规范候选版", "full_name": "MCP Specification 2026-07-28 RC", "category": "协议规范", "description": "MCP 2026年7月规范候选版发布：无状态协议核心、Extensions框架、Tasks、MCP Apps等重大更新", "protocol_version": "2026-07-28", "auth": "OAuth 2.1", "hosts": [ "Claude Desktop", "VS Code", "Cursor", "Gemini" ], "security": "企业级", "stars": 0, "maintainer": "Anthropic", "status": "active", "language": "N/A", "features": [ "无状态协议核心", "Extensions框架", "Tasks", "MCP Apps" ] }, { "id": "MCP-173", "name": "MCP企业部署指南 2026", "full_name": "MCP Enterprise Deployment Guide", "category": "企业部署", "description": "2026年MCP企业部署完整指南：JSON-RPC规范、OAuth 2.1安全、可观测性模式和部署路线图", "protocol_version": "2025-06-18+", "auth": "OAuth 2.1", "hosts": [ "Enterprise" ], "security": "企业级安全", "stars": 0, "maintainer": "社区", "status": "active", "language": "N/A", "features": [ "OAuth 2.1", "可观测性", "企业部署路线图" ] }, { "id": "MCP-174", "name": "MCP Extensions框架", "full_name": "MCP Extensions Framework", "category": "扩展框架", "description": "MCP Extensions框架允许社区构建可复用的协议扩展，支持自定义能力和工具集成", "protocol_version": "2026-07-28", "auth": "varies", "hosts": [ "兼容MCP客户端" ], "security": "扩展定义", "stars": 0, "maintainer": "Anthropic+社区", "status": "active", "language": "TypeScript/Python", "features": [ "可扩展架构", "社区扩展", "自定义工具" ] }, { "id": "MCP-175", "name": "MCP Apps", "full_name": "MCP Apps", "category": "应用框架", "description": "MCP Apps是MCP 2026规范引入的应用层概念，支持构建完整的MCP原生应用", "protocol_version": "2026-07-28", "auth": "varies", "hosts": [ "MCP兼容平台" ], "security": "应用级", "stars": 0, "maintainer": "Anthropic", "status": "active", "language": "Multi", "features": [ "应用层抽象", "原生MCP应用", "端到端工具链" ] }, { "id": "MCP-176", "name": "MCP Tasks", "full_name": "MCP Tasks", "category": "任务调度", "description": "MCP Tasks是2026规范引入的任务调度机制，支持异步任务和长运行操作", "protocol_version": "2026-07-28", "auth": "varies", "hosts": [ "MCP兼容客户端" ], "security": "任务级", "stars": 0, "maintainer": "Anthropic", "status": "active", "language": "Multi", "features": [ "异步任务", "长运行操作", "任务状态追踪" ] }, { "id": "MCP-177", "name": "MCP无状态协议核心", "full_name": "MCP Stateless Core", "category": "协议核心", "description": "MCP 2026规范将协议核心重构为无状态设计，简化扩展和部署，提高可伸缩性", "protocol_version": "2026-07-28", "auth": "varies", "hosts": [ "全平台" ], "security": "无状态安全", "stars": 0, "maintainer": "Anthropic", "status": "active", "language": "N/A", "features": [ "无状态设计", "高可伸缩", "简化部署" ] }, { "id": "MCP-178", "name": "MCP生态统计 2026", "full_name": "MCP Ecosystem Stats 2026", "category": "生态统计", "description": "2026年MCP采用统计：已验证服务器数量、GitHub生态信号、企业生产数据和集成覆盖", "protocol_version": "N/A", "auth": "N/A", "hosts": [ "N/A" ], "security": "N/A", "stars": 0, "maintainer": "社区", "status": "active", "language": "N/A", "features": [ "生态数据", "采用统计", "企业生产指标" ] }, { "id": "MCP-179", "name": "postgres-connector", "full_name": "PostgreSQL Database Connector", "category": "Database", "description": "Read/write access to PostgreSQL databases with query validation and parameterized statement support." }, { "id": "MCP-180", "name": "sqlite-tools", "full_name": "SQLite File System Connector", "category": "Database", "description": "Lightweight local SQLite database integration for schema inspection and data queries." }, { "id": "MCP-181", "name": "rest-api-client", "full_name": "Generic REST API Tool", "category": "API", "description": "Execute HTTP requests against arbitrary REST endpoints with configurable headers and authentication." }, { "id": "MCP-182", "name": "graphql-fetcher", "full_name": "GraphQL Schema Explorer", "category": "API", "description": "Introspect and query GraphQL APIs with schema discovery and typed response handling." }, { "id": "MCP-183", "name": "local-fs", "full_name": "Local File System Connector", "category": "File System", "description": "Safe read/write/list operations on local directories with path traversal protection." }, { "id": "MCP-184", "name": "s3-storage", "full_name": "AWS S3 Cloud Storage", "category": "Cloud Service", "description": "Upload, download, and list objects in Amazon S3 buckets with IAM credential support." }, { "id": "MCP-185", "name": "gcp-cloudsql", "full_name": "Google Cloud SQL Connector", "category": "Database", "description": "Connect to Google Cloud SQL instances for MySQL and PostgreSQL with managed authentication." }, { "id": "MCP-186", "name": "azure-blob", "full_name": "Azure Blob Storage Connector", "category": "Cloud Service", "description": "Manage Azure Blob Storage containers and blobs with SAS token or managed identity auth." }, { "id": "MCP-187", "name": "docker-manager", "full_name": "Docker Container Tool", "category": "Dev Tool", "description": "Inspect, start, stop, and list Docker containers and images via the local Docker daemon." }, { "id": "MCP-188", "name": "k8s-cluster", "full_name": "Kubernetes Cluster Connector", "category": "Dev Tool", "description": "Query pods, services, deployments, and events from a Kubernetes cluster using kubeconfig." }, { "id": "MCP-189", "name": "github-repo", "full_name": "GitHub Repository Tool", "category": "Dev Tool", "description": "List repos, read files, search issues, and manage pull requests via the GitHub API." }, { "id": "MCP-190", "name": "gitlab-project", "full_name": "GitLab Project Connector", "category": "Dev Tool", "description": "Access GitLab projects, read source files, and query merge requests programmatically." }, { "id": "MCP-191", "name": "mongodb-driver", "full_name": "MongoDB NoSQL Connector", "category": "Database", "description": "Query and aggregate data from MongoDB instances with connection string authentication." }, { "id": "MCP-192", "name": "redis-cache", "full_name": "Redis Cache Manager", "category": "Database", "description": "Read/write/cache keys in Redis with support for string, hash, and list data structures." }, { "id": "MCP-193", "name": "terraform-state", "full_name": "Terraform State Remote Backend", "category": "Dev Tool", "description": "Read and inspect Terraform state files stored in remote backends like S3 or GCS." }, { "id": "MCP-194", "name": "postgres-server", "full_name": "PostgreSQL Database Connector", "category": "Database Connector", "description": "Connects to PostgreSQL databases for querying, schema inspection, and transactional operations." }, { "id": "MCP-195", "name": "mysql-server", "full_name": "MySQL Database Connector", "category": "Database Connector", "description": "Provides read/write access to MySQL databases with support for complex queries and migrations." }, { "id": "MCP-196", "name": "sqlite-server", "full_name": "SQLite File-Based Database", "category": "Database Connector", "description": "Lightweight serverless SQLite connector for local development and small-scale data storage." }, { "id": "MCP-197", "name": "mongodb-server", "full_name": "MongoDB Document Store Connector", "category": "Database Connector", "description": "Connects to MongoDB clusters for document CRUD, aggregation pipelines, and indexing." }, { "id": "MCP-198", "name": "github-server", "full_name": "GitHub Repository & API Tool", "category": "API Tool", "description": "Interacts with GitHub REST/GraphQL APIs for issues, PRs, repos, and code review workflows." }, { "id": "MCP-199", "name": "gitlab-server", "full_name": "GitLab CI/CD & Repo Tool", "category": "API Tool", "description": "Manages GitLab projects, merge requests, CI pipelines, and issues via the GitLab API." }, { "id": "MCP-200", "name": "slack-server", "full_name": "Slack Messaging Connector", "category": "API Tool", "description": "Sends messages, manages channels, and integrates with Slack workspaces for team communication." }, { "id": "MCP-201", "name": "jira-server", "full_name": "Atlassian Jira Project Tracker", "category": "API Tool", "description": "Queries and updates Jira tickets, sprints, and boards for agile project management." }, { "id": "MCP-202", "name": "filesystem-server", "full_name": "Local File System Access", "category": "File System", "description": "Reads, writes, searches, and manipulates files on the local disk with path-safe operations." }, { "id": "MCP-203", "name": "gdrive-server", "full_name": "Google Drive Cloud Storage", "category": "Cloud Service", "description": "Accesses Google Drive for file upload, download, sharing, and folder management." }, { "id": "MCP-204", "name": "aws-s3-server", "full_name": "AWS S3 Object Storage", "category": "Cloud Service", "description": "Interacts with Amazon S3 buckets for storing and retrieving objects and managing permissions." }, { "id": "MCP-205", "name": "docker-server", "full_name": "Docker Container Management", "category": "Dev Tool", "description": "Manages Docker containers, images, and networks for local development and testing environments." }, { "id": "MCP-206", "name": "kubernetes-server", "full_name": "Kubernetes Cluster Orchestrator", "category": "Dev Tool", "description": "Deploys, scales, and monitors Kubernetes pods, services, and configurations remotely." }, { "id": "MCP-207", "name": "redis-server", "full_name": "Redis Cache & Session Store", "category": "Database Connector", "description": "Connects to Redis instances for caching, pub/sub messaging, and session data management." }, { "id": "MCP-208", "name": "vercel-server", "full_name": "Vercel Deployment Platform", "category": "Cloud Service", "description": "Triggers deployments, manages projects, and retrieves logs from the Vercel hosting platform." }, { "id": "MCP-209", "name": "postgres-connector", "full_name": "PostgreSQL Database Connector", "category": "Database Connector", "description": "Connects to PostgreSQL databases for querying, inserting, updating, and managing relational data with full transaction support." }, { "id": "MCP-210", "name": "mysql-toolkit", "full_name": "MySQL Database Toolkit", "category": "Database Connector", "description": "Provides read/write access to MySQL databases with schema inspection, query execution, and migration capabilities." }, { "id": "M211", "name": "redis-cache", "full_name": "Redis Cache & Session Store", "category": "Database Connector", "description": "Interacts with Redis instances for caching, session management, pub/sub messaging, and key-value operations." }, { "id": "MCP-212", "name": "mongo-db", "full_name": "MongoDB Document Store", "category": "Database Connector", "description": "Connects to MongoDB for document CRUD operations, aggregation pipelines, and index management." }, { "id": "MCP-213", "name": "sqlite-manager", "full_name": "SQLite Local Database Manager", "category": "Database Connector", "description": "Manages local SQLite databases for lightweight prototyping and embedded application storage." }, { "id": "MCP-214", "name": "rest-api-client", "full_name": "RESTful API Tool Client", "category": "API Tool", "description": "Executes HTTP requests against REST APIs with support for authentication headers, query params, and body payloads." }, { "id": "MCP-215", "name": "graphql-proxy", "full_name": "GraphQL Query Proxy", "category": "API Tool", "description": "Executes GraphQL queries and mutations against remote endpoints with schema introspection and variable support." }, { "id": "MCP-216", "name": "openapi-spec-gen", "full_name": "OpenAPI Specification Generator", "category": "API Tool", "description": "Generates OpenAPI 3.0/3.1 specifications from live API traffic or annotated source code." }, { "id": "MCP-217", "name": "swagger-ui-server", "full_name": "Swagger UI Interactive Server", "category": "API Tool", "description": "Serves interactive Swagger UI documentation for discovered APIs with inline try-it-out functionality." }, { "id": "MCP-218", "name": "webhook-dispatcher", "full_name": "Webhook Dispatcher Service", "category": "API Tool", "description": "Sends outbound webhook payloads with retry logic, signature signing, and delivery status tracking." }, { "id": "MCP-219", "name": "local-fs", "full_name": "Local File System Bridge", "category": "File System", "description": "Provides secure read/write/list/search operations on the local filesystem with path validation and permission checks." }, { "id": "MCP-220", "name": "s3-storage", "full_name": "Amazon S3 Object Storage", "category": "Cloud Service", "description": "Manages objects in Amazon S3 buckets including upload, download, listing, and lifecycle policy configuration." }, { "id": "MCP-221", "name": "gcs-bridge", "full_name": "Google Cloud Storage Bridge", "category": "Cloud Service", "description": "Interfaces with Google Cloud Storage for blob uploads, downloads, bucket management, and signed URL generation." }, { "id": "MCP-222", "name": "azure-blob", "full_name": "Azure Blob Storage Connector", "category": "Cloud Service", "description": "Connects to Azure Blob Storage for managing containers, uploading blobs, and retrieving access policies." }, { "id": "MCP-223", "name": "docker-manager", "full_name": "Docker Container Manager", "category": "Dev Tool", "description": "Controls Docker containers—building images, starting/stopping services, inspecting logs, and managing networks." }, { "id": "MCP-224", "name": "k8s-controller", "full_name": "Kubernetes Cluster Controller", "category": "Dev Tool", "description": "Interacts with Kubernetes clusters to manage pods, deployments, services, and configmaps via kubectl-like operations." }, { "id": "MCP-225", "name": "git-repo-tool", "full_name": "Git Repository Management Tool", "category": "Dev Tool", "description": "Performs git operations including clone, commit, push, diff, branch creation, and log inspection on local repositories." }, { "id": "MCP-226", "name": "terraform-plan", "full_name": "Terraform Infrastructure Planner", "category": "Dev Tool", "description": "Executes Terraform plan and apply workflows to provision and manage cloud infrastructure as code." }, { "id": "MCP-227", "name": "prometheus-metrics", "full_name": "Prometheus Metrics Collector", "category": "Dev Tool", "description": "Queries Prometheus for application metrics, generates alerts, and exports service health dashboards." }, { "id": "MCP-228", "name": "vault-secrets", "full_name": "HashiCorp Vault Secrets Manager", "category": "Cloud Service", "description": "Reads and writes secrets from HashiCorp Vault with dynamic credential rotation and audit logging." }, { "id": "MCP-224", "name": "PostgreSQL Connector", "full_name": "PostgreSQL Database MCP Server", "category": "database", "description": "Connects to PostgreSQL databases for querying, schema inspection, and transaction management." }, { "id": "MCP-225", "name": "MySQL Bridge", "full_name": "MySQL Database MCP Server", "category": "database", "description": "Enables read/write operations on MySQL instances with support for stored procedures and migrations." }, { "id": "MCP-226", "name": "MongoDB Atlas", "full_name": "MongoDB Atlas Cloud Database Server", "category": "database", "description": "Provides access to MongoDB Atlas clusters for document-based data operations and aggregation pipelines." }, { "id": "MCP-227", "name": "REST API Client", "full_name": "Generic REST API Tool Server", "category": "api", "description": "Executes HTTP requests against RESTful APIs with configurable headers, auth, and response parsing." }, { "id": "MCP-228", "name": "GraphQL Gateway", "full_name": "GraphQL Schema Explorer Server", "category": "api", "description": "Introspects and executes GraphQL queries against remote or local schema endpoints." }, { "id": "MCP-229", "name": "Local File System", "full_name": "Local Filesystem MCP Server", "category": "filesystem", "description": "Reads, writes, and navigates the local filesystem with safe path resolution and filtering." }, { "id": "MCP-230", "name": "S3 Storage", "full_name": "AWS S3 Cloud Storage Server", "category": "cloud", "description": "Manages objects in Amazon S3 buckets including upload, download, list, and delete operations." }, { "id": "MCP-231", "name": "Azure Blob", "full_name": "Azure Blob Storage MCP Server", "category": "cloud", "description": "Interacts with Azure Blob Storage for managing containers, blobs, and metadata." }, { "id": "MCP-232", "name": "Docker Engine", "full_name": "Docker Container Management Server", "category": "devtool", "description": "Controls Docker containers, images, and networks via the Docker daemon API." }, { "id": "MCP-233", "name": "Git Repo Manager", "full_name": "Git Repository MCP Server", "category": "devtool", "description": "Performs Git operations including clone, commit, branch, diff, and log retrieval." }, { "id": "MCP-234", "name": "Redis Cache", "full_name": "Redis In-Memory Store Server", "category": "database", "description": "Connects to Redis for key-value caching, pub/sub messaging, and session management." }, { "id": "MCP-235", "name": "Slack Bot", "full_name": "Slack Workspace Integration Server", "category": "api", "description": "Reads and sends messages in Slack channels and direct messages via the Slack API." }, { "id": "MCP-236", "name": "Elasticsearch", "full_name": "Elasticsearch Search Engine Server", "category": "database", "description": "Indexes and searches documents using Elasticsearch with full-text query capabilities." }, { "id": "MCP-237", "name": "Kubernetes Controller", "full_name": "K8s Cluster Management Server", "category": "cloud", "description": "Manages Kubernetes pods, deployments, services, and configmaps across clusters." }, { "id": "MCP-238", "name": "GitHub Actions", "full_name": "GitHub CI/CD Pipeline Server", "category": "devtool", "description": "Triggers and monitors GitHub Actions workflows for continuous integration and deployment." }, { "id": "MCP-179", "name": "Filesystem Server", "full_name": "Model Context Protocol Filesystem Server", "category": "System", "description": "Provides read and write access to the local file system with secure path resolution." }, { "id": "MCP-180", "name": "Postgres Server", "full_name": "PostgreSQL Database Server", "category": "Database", "description": "Enables querying and managing PostgreSQL databases with transaction support." }, { "id": "MCP-181", "name": "GitHub Server", "full_name": "GitHub Integration Server", "category": "DevOps", "description": "Connects to GitHub API for repository management, issue tracking, and pull request operations." }, { "id": "MCP-182", "name": "GitLab Server", "full_name": "GitLab Integration Server", "category": "DevOps", "description": "Integrates with GitLab for project management, CI/CD pipeline monitoring, and code review." }, { "id": "MCP-183", "name": "Slack Server", "full_name": "Slack Messaging Server", "category": "Communication", "description": "Facilitates sending messages, channels management, and notifications via Slack API." }, { "id": "MCP-184", "name": "Google Drive Server", "full_name": "Google Drive File Server", "category": "Storage", "description": "Provides access to Google Drive for uploading, downloading, and organizing cloud files." }, { "id": "MCP-185", "name": "Sentry Server", "full_name": "Sentry Error Tracking Server", "category": "Monitoring", "description": "Connects to Sentry for real-time error monitoring, alerting, and stack trace analysis." }, { "id": "MCP-186", "name": "Puppeteer Server", "full_name": "Browser Automation Server", "category": "Web", "description": "Uses Puppeteer for headless browser automation, screenshot capture, and web scraping." }, { "id": "MCP-187", "name": "Linear Server", "full_name": "Linear Project Management Server", "category": "Project Management", "description": "Integrates with Linear for issue tracking, sprint planning, and team workflow management." }, { "id": "MCP-188", "name": "Redis Server", "full_name": "Redis Cache Server", "category": "Database", "description": "Provides fast key-value store operations including caching, pub/sub, and session management." }, { "id": "MCP-189", "name": "Elasticsearch Server", "full_name": "Elasticsearch Search Server", "category": "Search", "description": "Enables full-text search, indexing, and analytics on large datasets via Elasticsearch." }, { "id": "MCP-190", "name": "Docker Server", "full_name": "Docker Container Server", "category": "DevOps", "description": "Manages Docker containers, images, and networks for containerized application workflows." }, { "id": "MCP-191", "name": "Notion Server", "full_name": "Notion Workspace Server", "category": "Productivity", "description": "Connects to Notion API for reading, creating, and updating pages and databases." }, { "id": "MCP-192", "name": "AWS S3 Server", "full_name": "Amazon S3 Storage Server", "category": "Cloud", "description": "Provides object storage operations for Amazon S3 including bucket management and file transfers." }, { "id": "MCP-193", "name": "Terminal Server", "full_name": "Interactive Terminal Server", "category": "System", "description": "Executes shell commands securely with output streaming and timeout controls." }, { "id": "MCP-194", "name": "File System Server", "full_name": "Model Context Protocol File System Server", "category": "System", "description": "Provides secure file system operations including reading, writing, and listing files within designated directories." }, { "id": "MCP-195", "name": "PostgreSQL Server", "full_name": "Model Context Protocol PostgreSQL Database Server", "category": "Database", "description": "Enables read-only querying of PostgreSQL databases with support for complex SQL queries and schema inspection." }, { "id": "MCP-196", "name": "GitHub Server", "full_name": "Model Context Protocol GitHub Integration Server", "category": "Development", "description": "Connects to GitHub API to manage repositories, issues, pull requests, and code reviews programmatically." }, { "id": "MCP-197", "name": "Memory Store Server", "full_name": "Model Context Protocol Persistent Memory Store Server", "category": "Storage", "description": "Implements a key-value memory store for persisting context and conversation state across sessions." }, { "id": "MCP-198", "name": "Web Browser Server", "full_name": "Model Context Protocol Web Browser Automation Server", "category": "Automation", "description": "Controls headless browser instances for web scraping, testing, and interactive web page navigation." }, { "id": "MCP-199", "name": "Puppeteer Server", "full_name": "Model Context Protocol Puppeteer Automation Server", "category": "Automation", "description": "Leverages Puppeteer for advanced browser automation, screenshot capture, and DOM interaction capabilities." }, { "id": "MCP-200", "name": "Sentry Server", "full_name": "Model Context Protocol Sentry Error Monitoring Server", "category": "Observability", "description": "Integrates with Sentry to fetch error reports, trace stack details, and monitor application health metrics." }, { "id": "MCP-201", "name": "Git Server", "full_name": "Model Context Protocol Local Git Repository Server", "category": "Development", "description": "Interacts with local Git repositories for commit history, branch management, diff viewing, and log analysis." }, { "id": "MCP-202", "name": "Slack Server", "full_name": "Model Context Protocol Slack Messaging Server", "category": "Communication", "description": "Connects to Slack workspace to read channels, send messages, and retrieve conversation history." }, { "id": "MCP-203", "name": "Linear Server", "full_name": "Model Context Protocol Linear Project Management Server", "category": "Project Management", "description": "Interfaces with Linear API for managing issues, projects, sprints, and team workflows." }, { "id": "MCP-204", "name": "Vector DB Server", "full_name": "Model Context Protocol Vector Database Retrieval Server", "category": "AI/ML", "description": "Provides semantic search and similarity retrieval over vector embeddings stored in supported vector databases." }, { "id": "MCP-205", "name": "Time Tracking Server", "full_name": "Model Context Protocol Time Intelligence Server", "category": "Productivity", "description": "Offers timezone conversion, calendar integration, and time-based scheduling utilities for task planning." }, { "id": "MCP-206", "name": "Docker Server", "full_name": "Model Context Protocol Docker Container Management Server", "category": "DevOps", "description": "Manages Docker containers including listing images, starting/stopping services, and inspecting container logs." }, { "id": "MCP-207", "name": "LLM Gateway Server", "full_name": "Model Context Protocol Large Language Model Gateway Server", "category": "AI/ML", "description": "Acts as an abstraction layer for multiple LLM providers, standardizing prompt templates and response formats." }, { "id": "MCP-208", "name": "Search Engine Server", "full_name": "Model Context Protocol Web Search Aggregation Server", "category": "Information Retrieval", "description": "Aggregates results from multiple search engines and provides deduplicated, ranked search answer summaries." } ], "agent_sdks": [ { "id": "ASDK-001", "name": "Vercel AI SDK", "vendor": "Vercel", "language": "TypeScript", "type": "AI SDK", "version": "4.x", "license": "Apache-2.0", "features": [ "统一API", "Streaming", "Tool Calling", "RAG", "Middleware", "React/Svelte/Vue组件" ], "pros": [ "前端集成最好", "SSR友好", "框架无关" ], "cons": [ "JS/TS only", "无内置多Agent" ], "maturity": "生产可用", "docs": "https://sdk.vercel.ai/docs" }, { "id": "ASDK-002", "name": "LangChain", "vendor": "LangChain", "language": "Python/JS", "type": "AI框架", "version": "0.3+", "license": "MIT", "features": [ "Chain编排", "Memory", "Retrieval", "Agent", "Callbacks", "200+集成" ], "pros": [ "集成最多", "社区最大", "文档丰富" ], "cons": [ "抽象层太多", "性能差", "调试难" ], "maturity": "生产可用", "docs": "https://python.langchain.com/docs" }, { "id": "ASDK-003", "name": "LlamaIndex", "vendor": "LlamaIndex", "language": "Python/TS", "type": "RAG框架", "version": "0.12+", "license": "Apache-2.0", "features": [ "数据摄取", "索引构建", "查询引擎", "Agent", "Workflow" ], "pros": [ "RAG最强", "数据连接器多", "灵活" ], "cons": [ "RAG外功能弱", "学习曲线" ], "maturity": "生产可用", "docs": "https://docs.llamaindex.ai" }, { "id": "ASDK-004", "name": "Haystack", "vendor": "deepset", "language": "Python", "type": "RAG/搜索框架", "version": "2.x", "license": "Apache-2.0", "features": [ "Pipeline", "组件化", "RAG", "Agent", "搜索" ], "pros": [ "组件化设计", "生产级", "可测试" ], "cons": [ "社区比LangChain小", "文档一般" ], "maturity": "生产可用", "docs": "https://docs.haystack.deepset.ai" }, { "id": "ASDK-005", "name": "Semantic Kernel", "vendor": "Microsoft", "language": "C#/Python/Java", "type": "AI SDK", "version": "1.x", "license": "MIT", "features": [ "Plugin系统", "Planner", "Memory", "Connector" ], "pros": [ "企业级", ".NET生态", "微软背书" ], "cons": [ "C#优先", "Python版落后", "社区小" ], "maturity": "生产可用", "docs": "https://learn.microsoft.com/en-us/semantic-kernel" }, { "id": "ASDK-006", "name": "Rig", "vendor": "0xPlay", "language": "Rust", "type": "AI SDK", "version": "0.8+", "license": "MIT", "features": [ "LLM调用", "Agent", "Tool", "RAG", "Embedding" ], "pros": [ "Rust性能", "类型安全", "轻量" ], "cons": [ "Rust only", "生态小", "文档少" ], "maturity": "早期", "docs": "https://rig.rs/docs" }, { "id": "ASDK-007", "name": "Portkey AI Gateway", "vendor": "Portkey", "language": "TypeScript", "type": "AI网关", "version": "1.x", "license": "MIT", "features": [ "多模型路由", "Fallback", "缓存", "限流", "Observability" ], "pros": [ "模型无关", "生产监控", "成本控制" ], "cons": [ "网关层开销", "依赖第三方" ], "maturity": "生产可用", "docs": "https://portkey.ai/docs" }, { "id": "ASDK-008", "name": "Helicone", "vendor": "Helicone", "language": "TypeScript", "type": "AI可观测性", "version": "2.x", "license": "Apache-2.0", "features": [ "日志", "监控", "缓存", "速率限制", "实验" ], "pros": [ "开箱即用", "多模型支持", "免费层" ], "cons": [ "需代理层", "延迟增加" ], "maturity": "生产可用", "docs": "https://docs.helicone.ai" }, { "id": "SDK-104", "name": "LangChain", "vendor": "LangChain Inc.", "language": "Python/JavaScript", "type": "Orchestration Framework" }, { "id": "SDK-105", "name": "AutoGen", "vendor": "Microsoft", "language": "Python", "type": "Multi-Agent Framework" }, { "id": "SDK-106", "name": "CrewAI", "vendor": "CrewAI Inc.", "language": "Python", "type": "Role-Based Framework" }, { "id": "SDK-107", "name": "LangGraph", "vendor": "LangChain", "language": "Python/JavaScript", "type": "State Machine" }, { "id": "SDK-108", "name": "LlamaIndex", "vendor": "LlamaIndex Inc.", "language": "Python", "type": "Data Framework" }, { "id": "SDK-109", "name": "OpenAI Agents SDK", "vendor": "OpenAI", "language": "Python/TypeScript", "type": "Native SDK" }, { "id": "SDK-110", "name": "Google ADK", "vendor": "Google", "language": "Python/TypeScript", "type": "Agent Development Kit" }, { "id": "SDK-111", "name": "AWS Bedrock Agents", "vendor": "Amazon", "language": "Python/Java", "type": "Cloud Service SDK" }, { "id": "SDK-112", "name": "Microsoft AutoGen Core", "vendor": "Microsoft", "language": "Python", "type": "Core Runtime" }, { "id": "SDK-113", "name": "Haystack", "vendor": "Deepset", "language": "Python", "type": "RAG Framework" }, { "id": "SDK-114", "name": "Semantic Kernel", "vendor": "Microsoft", "language": "C#/Python/Java", "type": "SDK" }, { "id": "SDK-115", "name": "Dify", "vendor": "Dify.ai", "language": "Python/TypeScript", "type": "LLM App Platform" }, { "id": "SDK-116", "name": "FlowiseAI", "vendor": "FlowiseAI", "language": "TypeScript", "type": "Drag-and-Drop Builder" }, { "id": "SDK-117", "name": "Vercel AI SDK", "vendor": "Vercel", "language": "TypeScript", "type": "Streaming UI SDK" }, { "id": "SDK-118", "name": "Smithery", "vendor": "Smithery.ai", "language": "Language Agnostic", "type": "Agent Registry SDK" }, { "id": "SDK-119", "name": "AgentKit Core", "vendor": "NeuroSoft", "language": "Python", "type": "Framework" }, { "id": "SDK-120", "name": "AutoFlow AI", "vendor": "StreamLine Tech", "language": "JavaScript", "type": "Library" }, { "id": "SDK-121", "name": "MindBridge SDK", "vendor": "Cognitive Labs", "language": "Go", "type": "Toolkit" }, { "id": "SDK-122", "name": "Synapse Agent", "vendor": "DeepLink AI", "language": "Rust", "type": "Framework" }, { "id": "SDK-123", "name": "NexusAI Connect", "vendor": "Quantum Minds", "language": "Python", "type": "Library" }, { "id": "SDK-124", "name": "BotForge SDK", "vendor": "Industrial AI", "language": "Java", "type": "Toolkit" }, { "id": "SDK-125", "name": "EchoAgent", "vendor": "Resonance Tech", "language": "C#", "type": "Framework" }, { "id": "SDK-126", "name": "Vortex AI Kit", "vendor": "Spin Dynamics", "language": "Python", "type": "Library" }, { "id": "SDK-127", "name": "OmniAgent SDK", "vendor": "Universal Logic", "language": "TypeScript", "type": "Toolkit" }, { "id": "SDK-128", "name": "PrismAI Engine", "vendor": "Spectrum Systems", "language": "Rust", "type": "Framework" }, { "id": "SDK-129", "name": "FluxAgent", "vendor": "Dynamic Flow", "language": "Go", "type": "Library" }, { "id": "SDK-130", "name": "CoreBrain SDK", "vendor": "Central Intel", "language": "Python", "type": "Toolkit" }, { "id": "SDK-131", "name": "AgentWave", "vendor": "Tidal AI", "language": "JavaScript", "type": "Framework" }, { "id": "SDK-132", "name": "LogicLoop SDK", "vendor": "Infinite Loops", "language": "Java", "type": "Library" }, { "id": "SDK-133", "name": "AetherAgent", "vendor": "Sky High Tech", "language": "C++", "type": "Toolkit" } ], "protocols": [ { "id": "PROTO-001", "name": "MCP (Model Context Protocol)", "vendor": "Anthropic", "version": "2026-07-28-rc", "layer": "Agent-Tool", "description": "让AI Agent连接外部工具和数据源的开放协议，2026年最大版本更新", "architecture": "Client-Server", "transport": "stdio/SSE/StreamableHTTP", "security": "进程隔离/权限控制/OAuth", "adoption": "广泛（13000+ MCP Server）", "github_stars": "30000+", "status": "事实标准" }, { "id": "PROTO-002", "name": "A2A (Agent-to-Agent)", "vendor": "Google", "version": "0.1", "layer": "Agent-Agent", "description": "让不同框架的AI Agent互相通信和协作", "architecture": "P2P/Hub", "transport": "HTTP/JSON-RPC", "security": "Agent Card认证", "adoption": "快速增长(50+合作伙伴)", "github_stars": "5000+", "status": "快速增长" }, { "id": "PROTO-003", "name": "ACP (Agent Communication Protocol)", "vendor": "IBM/BEA", "version": "0.1", "layer": "Agent-Agent", "description": "IBM主导的Agent通信协议", "architecture": "Hub-Spoke", "transport": "HTTP/gRPC", "security": "企业级安全", "adoption": "IBM生态", "github_stars": "500", "status": "早期" }, { "id": "PROTO-004", "name": "OpenClaw Skill Protocol", "vendor": "OpenClaw", "version": "1.0", "layer": "Agent-Skill", "description": "Agent技能描述和分发协议", "architecture": "Registry", "transport": "npm/git", "security": "沙箱执行/权限声明", "adoption": "OpenClaw生态", "github_stars": "2000", "status": "活跃" }, { "id": "PROTO-005", "name": "LSP (Language Server Protocol)", "vendor": "Microsoft", "version": "3.17", "layer": "Editor-Language", "description": "编辑器与语言服务通信协议", "architecture": "Client-Server", "transport": "stdio/HTTP", "security": "进程隔离", "adoption": "广泛", "github_stars": "11000", "status": "成熟标准" }, { "id": "PROTO-006", "name": "AG-UI (Agent User Interaction Protocol)", "vendor": "CopilotKit", "version": "0.1", "layer": "Agent-UI", "description": "Agent与前端UI交互的开放协议，定义Agent事件流到UI组件的映射", "architecture": "Event Stream", "transport": "SSE/WebSocket", "security": "CORS/认证", "adoption": "CopilotKit生态", "github_stars": "8000+", "status": "快速增长" }, { "id": "PROTO-007", "name": "ANP (Agent Network Protocol)", "vendor": "ANP社区", "version": "0.2", "layer": "Agent-Network", "description": "去中心化Agent网络协议，支持Agent发现和互操作", "architecture": "P2P/DID", "transport": "HTTP/DID Comm", "security": "DID认证/加密", "adoption": "W3C DID生态", "github_stars": "3000", "status": "活跃" }, { "id": "PROTO-008", "name": "OpenAPI/Swagger", "vendor": "Linux Foundation", "version": "3.1", "layer": "API-Description", "description": "REST API描述标准，MCP Server可从OpenAPI自动生成", "architecture": "Specification", "transport": "HTTP", "security": "OAuth/API Key", "adoption": "广泛", "github_stars": "29000", "status": "成熟标准" }, { "id": "PROTO-009", "name": "gRPC", "vendor": "Google", "version": "1.6x", "layer": "RPC", "description": "高性能RPC框架，Agent间低延迟通信", "architecture": "Client-Server/Streaming", "transport": "HTTP/2", "security": "TLS/Token", "adoption": "广泛", "github_stars": "41000", "status": "成熟标准" }, { "id": "PROTO-010", "name": "GraphQL", "vendor": "GraphQL Foundation", "version": "October 2021", "layer": "API-Query", "description": "灵活的API查询语言，Agent按需获取数据", "architecture": "Client-Server", "transport": "HTTP/WebSocket", "security": "查询深度限制/认证", "adoption": "广泛", "github_stars": "14000", "status": "成熟标准" }, { "id": "PROTO-011", "name": "WASM Component Model", "vendor": "Bytecode Alliance", "version": "0.2", "layer": "Agent-Sandbox", "description": "WebAssembly组件模型，Agent安全沙箱执行", "architecture": "Component", "transport": "WASM Runtime", "security": "沙箱隔离/能力控制", "adoption": "增长中", "github_stars": "1000", "status": "活跃" }, { "id": "PROTO-012", "name": "JSON-RPC 2.0", "vendor": "JSON-RPC社区", "version": "2.0", "layer": "RPC", "description": "轻量级远程过程调用协议，MCP底层传输之一", "architecture": "Client-Server", "transport": "HTTP/WebSocket/stdio", "security": "传输层安全", "adoption": "广泛", "github_stars": "N/A", "status": "成熟标准" }, { "id": "PROTO-013", "name": "Server-Sent Events (SSE)", "vendor": "W3C", "version": "Living Standard", "layer": "Transport", "description": "服务器推送事件流，MCP远程传输方式", "architecture": "Server-Push", "transport": "HTTP", "security": "CORS/认证", "adoption": "广泛", "github_stars": "N/A", "status": "成熟标准" }, { "id": "PROTO-014", "name": "DAP (Debug Adapter Protocol)", "vendor": "Microsoft", "version": "1.6x", "layer": "Agent-Debug", "description": "调试适配器协议，Agent调试和可观测性", "architecture": "Client-Server", "transport": "stdio/HTTP", "security": "进程隔离", "adoption": "IDE广泛", "github_stars": "2000", "status": "成熟" }, { "id": "PROTO-015", "name": "OpenTelemetry", "vendor": "CNCF", "version": "1.x", "layer": "Observability", "description": "可观测性标准，Agent tracing/metrics/logs", "architecture": "Collector/Export", "transport": "OTLP/gRPC/HTTP", "security": "TLS/Token", "adoption": "广泛", "github_stars": "4500", "status": "事实标准" }, { "id": "PROTO-016", "name": "FHIR (Healthcare API)", "vendor": "HL7", "version": "R4B/R5", "layer": "Domain-API", "description": "医疗健康数据交换标准，医疗Agent场景", "architecture": "RESTful", "transport": "HTTP", "security": "OAuth2/SMART", "adoption": "医疗行业广泛", "github_stars": "2000", "status": "行业标准" }, { "id": "PROTO-017", "name": "Stainless API Standard", "vendor": "Stainless", "version": "1.0", "layer": "API-SDK-Gen", "description": "API到SDK自动生成标准，MCP Server SDK生成", "architecture": "Generator", "transport": "HTTP", "security": "规范定义", "adoption": "增长中", "github_stars": "3000", "status": "活跃" }, { "id": "PROTO-018", "name": "Semantic Protocol (SP)", "vendor": "W3C", "version": "Draft", "layer": "Agent-Semantic", "description": "语义网协议栈，Agent知识表示和推理", "architecture": "RDF/OWL/SPARQL", "transport": "HTTP", "security": "访问控制", "adoption": "知识图谱领域", "github_stars": "N/A", "status": "标准草案" }, { "id": "PROTO-019", "name": "OAuth 2.1 / DPoP", "vendor": "IETF", "version": "RFC 9449", "layer": "Auth", "description": "Agent认证和授权标准，MCP远程安全基础", "architecture": "Token-based", "transport": "HTTP", "security": "PKCE/DPoP", "adoption": "广泛", "github_stars": "N/A", "status": "成熟标准" }, { "id": "PROTO-020", "name": "Tool Call Protocol (OpenAI)", "vendor": "OpenAI", "version": "2024-12", "layer": "Agent-Tool", "description": "OpenAI定义的函数调用协议，事实上的工具调用标准", "architecture": "Request-Response", "transport": "HTTP API", "security": "API Key", "adoption": "广泛（OpenAI兼容模型）", "github_stars": "N/A", "status": "事实标准" }, { "id": "PROTO-021", "name": "Agent Gateway Protocol (AGP)", "vendor": "社区", "version": "0.1", "layer": "Agent-Gateway", "description": "Agent网关协议，统一多Agent系统的入口和路由", "architecture": "Gateway", "transport": "HTTP/gRPC", "security": "mTLS/RBAC", "adoption": "早期", "github_stars": "1500", "status": "早期" }, { "id": "PROTO-022", "name": "x402 (HTTP 402 Payment Protocol)", "vendor": "Coinbase/社区", "version": "1.0", "layer": "Agent-Payment", "description": "基于HTTP 402状态码的机器对机器微支付协议，已处理7500万+交易", "architecture": "HTTP-native", "transport": "HTTP/HTTPS", "security": "加密签名/区块链", "adoption": "快速增长（75M+交易/月）", "github_stars": "5000+", "status": "生产可用" }, { "id": "PROTO-023", "name": "AP2 (Agent Payment Protocol)", "vendor": "Google", "version": "0.1", "layer": "Agent-Payment", "description": "Google推出的Agent预授权支付协议，60+合作组织", "architecture": "Pre-authorized", "transport": "HTTP/gRPC", "security": "OAuth/加密", "adoption": "早期（60+合作伙伴）", "github_stars": "N/A", "status": "早期" }, { "id": "PROTO-024", "name": "ACP (Agent Commerce Protocol)", "vendor": "Stripe", "version": "1.0", "layer": "Agent-Payment", "description": "Stripe推出的Agent商务协议，已集成到ChatGPT", "architecture": "Commerce-focused", "transport": "HTTPS", "security": "Stripe安全", "adoption": "ChatGPT集成", "github_stars": "N/A", "status": "生产可用" }, { "id": "PROTO-025", "name": "UCP (Universal Commerce Protocol)", "vendor": "Google", "version": "0.1", "layer": "Agent-Payment", "description": "Google推出的通用商务协议，用于Agent间交易", "architecture": "Universal", "transport": "HTTP/gRPC", "security": "加密", "adoption": "早期", "github_stars": "N/A", "status": "早期" }, { "id": "PROTO-027", "name": "A2UI (Agent-to-UI Protocol)", "vendor": "社区", "version": "0.1", "layer": "Agent-UI", "description": "Agent到用户界面的直接交互协议", "architecture": "Component-based", "transport": "WebSocket", "security": "Sandbox", "adoption": "早期", "github_stars": "500", "status": "早期" }, { "id": "PROTO-028", "name": "Web Bot Auth", "vendor": "IETF/社区", "version": "draft", "layer": "Agent-Auth", "description": "Web机器人身份认证标准，定义Agent在Web上的身份验证", "architecture": "Certificate-based", "transport": "HTTPS", "security": "PKI/OAuth", "adoption": "标准化中", "github_stars": "N/A", "status": "草案" }, { "id": "PROTO-029", "name": "MCP 2026-07-28 RC", "vendor": "Linux Foundation/MCP", "version": "2026-07-28-rc", "layer": "Agent-Tool", "description": "MCP协议2026年重大更新，引入无状态协议、Tasks原语、MCP Apps", "architecture": "Stateless Client-Server", "transport": "StreamableHTTP/SSE/stdio", "security": "OAuth 2.1/权限控制/审计日志", "adoption": "即将发布", "github_stars": "30000+", "status": "RC" }, { "id": "PROTO-030", "name": "MCP Registry Protocol", "vendor": "Linux Foundation/MCP", "version": "0.1", "layer": "Agent-Discovery", "description": "MCP Server注册与发现协议，用于Server目录和安全验证", "architecture": "Registry", "transport": "HTTPS", "security": "TLS/签名验证", "adoption": "预览版", "github_stars": "1500", "status": "预览" }, { "id": "PROTO-031", "name": "MCP Tasks Protocol", "vendor": "Linux Foundation/MCP", "version": "draft", "layer": "Agent-Task", "description": "MCP 2026路线图新增Tasks原语，支持Agent自主通信", "architecture": "Task-oriented", "transport": "StreamableHTTP", "security": "OAuth/权限", "adoption": "规划中", "github_stars": "N/A", "status": "规划" }, { "id": "PROTO-032", "name": "A2A v1", "vendor": "Google", "version": "1.0", "layer": "Agent-Agent", "description": "Agent间通信协议正式版，IBM ACP已合并入A2A", "architecture": "P2P/Hub", "transport": "HTTP/gRPC", "security": "mTLS/OAuth", "adoption": "广泛（IBM合并）", "github_stars": "15000+", "status": "正式版" }, { "id": "PROTO-033", "name": "Agent Pay (Mastercard)", "vendor": "Mastercard", "version": "1.0", "layer": "Agent-Payment", "description": "Mastercard Agent Pay协议，支持AI Agent安全支付", "architecture": "Payment Gateway", "transport": "HTTPS", "security": "Token化/3DS", "adoption": "发布", "github_stars": "N/A", "status": "生产可用" }, { "id": "PROTO-034", "name": "Visa Trusted Agent", "vendor": "Visa", "version": "1.0", "layer": "Agent-Payment", "description": "Visa可信Agent支付协议", "architecture": "Trusted Execution", "transport": "HTTPS", "security": "Token化/生物识别", "adoption": "发布", "github_stars": "N/A", "status": "生产可用" }, { "id": "PROTO-035", "name": "llms.txt Standard", "vendor": "社区", "version": "0.3", "layer": "Agent-Content", "description": "为LLM Agent提供网站内容摘要的标准格式", "architecture": "File-based", "transport": "HTTPS", "security": "CORS", "adoption": "广泛", "github_stars": "4000+", "status": "事实标准" }, { "id": "PROTO-036", "name": "Agent Identity Protocol (AIP)", "vendor": "社区", "version": "0.1", "layer": "Agent-Identity", "description": "定义AI Agent身份标识和验证的协议", "architecture": "DID-based", "transport": "HTTPS", "security": "DID/VC", "adoption": "早期", "github_stars": "200", "status": "草案" }, { "id": "PROTO-037", "name": "AGORA Protocol", "vendor": "Oxford University", "version": "0.1", "layer": "Agent-Agent", "description": "牛津大学提出的自然语言动态协议设计，Agent自适应任务协商，无需预定义协议", "architecture": "Natural Language Protocol", "transport": "HTTP/LLM", "security": "协商式安全", "adoption": "研究阶段", "github_stars": "500", "status": "研究" }, { "id": "PROTO-038", "name": "APEX Standard", "vendor": "APEX Community", "version": "0.1", "layer": "Agent-Interaction", "description": "开放Agent交互协议标准，定义Agent间发现、协商和执行的通用框架", "architecture": "Registry + P2P", "transport": "HTTP/WebSocket", "security": "OAuth/签名", "adoption": "早期", "github_stars": "300", "status": "草案" }, { "id": "PROTO-039", "name": "MCP Streamable HTTP", "vendor": "Linux Foundation/MCP", "version": "2026-07-28", "layer": "Agent-Transport", "description": "MCP 2026新增StreamableHTTP传输方式，替代SSE，支持无状态请求", "architecture": "Stateless HTTP", "transport": "HTTP/HTTPS", "security": "OAuth 2.1/CORS", "adoption": "即将发布", "github_stars": "N/A", "status": "RC" }, { "id": "PROTO-040", "name": "MCP Apps Protocol", "vendor": "Linux Foundation/MCP", "version": "draft", "layer": "Agent-App", "description": "MCP 2026路线图新增MCP Apps概念，定义可组合的Agent应用协议", "architecture": "Composable App", "transport": "StreamableHTTP", "security": "OAuth/权限", "adoption": "规划中", "github_stars": "N/A", "status": "规划" }, { "id": "PROTO-041", "name": "Agent Trust Protocol (ATP)", "vendor": "社区", "version": "0.1", "layer": "Agent-Trust", "description": "定义Agent信任评估和声誉系统的协议，支持跨平台Agent可信度验证", "architecture": "Reputation-based", "transport": "HTTPS", "security": "PKI/签名验证", "adoption": "早期", "github_stars": "200", "status": "草案" }, { "id": "PROTO-042", "name": "MCP 2026-07-28 RC (Stateless)", "vendor": "Linux Foundation/MCP", "version": "2026-07-28-rc", "layer": "Agent-Tool", "description": "MCP 2026最大版本更新，引入无状态协议核心、StreamableHTTP传输、Tasks原语、MCP Apps概念", "architecture": "Stateless Client-Server", "transport": "StreamableHTTP/SSE/stdio", "security": "OAuth 2.1/权限控制/审计日志", "adoption": "即将发布（RC阶段）", "github_stars": "30000+", "status": "RC", "sources": [ { "source_type": "web", "source_credibility": "A", "collected_at": "2026-06-01T13:48:20.198Z" } ] }, { "id": "PROTO-043", "name": "AGP (Agent Gateway Protocol)", "vendor": "社区", "version": "0.2", "layer": "Agent-Gateway", "description": "Agent网关协议升级版，统一多Agent系统入口、路由和安全策略", "architecture": "Gateway + Mesh", "transport": "HTTP/gRPC/WS", "security": "mTLS/RBAC/审计", "adoption": "增长中", "github_stars": "2000", "status": "活跃", "sources": [ { "source_type": "web", "source_credibility": "B", "collected_at": "2026-06-01T13:48:20.198Z" } ] }, { "id": "PROTO-044", "name": "MPP (Model Payment Protocol)", "vendor": "社区", "version": "0.1", "layer": "Agent-Payment", "description": "模型支付协议，定义AI模型调用和支付的标准化流程", "architecture": "Request-Payment-Response", "transport": "HTTPS", "security": "加密签名/区块链", "adoption": "早期", "github_stars": "300", "status": "草案", "sources": [ { "source_type": "web", "source_credibility": "C", "collected_at": "2026-06-01T13:48:20.198Z" } ] }, { "id": "PROTO-045", "name": "Agent Stack Protocol Suite", "vendor": "多厂商", "version": "2026", "layer": "Agent-Full-Stack", "description": "2026年Agent协议栈：MCP + A2A + x402 + AP2 + Web Bot Auth + llms.txt，形成完整Agent通信标准", "architecture": "Layered Stack", "transport": "Multiple", "security": "多层安全", "adoption": "行业趋势", "github_stars": "N/A", "status": "趋势", "sources": [ { "source_type": "web", "source_credibility": "A", "collected_at": "2026-06-01T13:48:20.198Z" } ] }, { "id": "PROTO-046", "name": "UCP (Universal Commerce Protocol) v2", "vendor": "Google", "version": "0.2", "layer": "Agent-Payment", "description": "Google通用商务协议升级版，与AP2合并，支持Agent间复杂交易", "architecture": "Universal + Pre-authorized", "transport": "HTTP/gRPC", "security": "OAuth/加密/审计", "adoption": "早期（60+合作伙伴）", "github_stars": "N/A", "status": "早期", "sources": [ { "source_type": "web", "source_credibility": "A", "collected_at": "2026-06-01T13:48:20.198Z" } ] }, { "id": "PROTO-047", "name": "TAP (Tool Access Protocol)", "type": "agent_tool_protocol", "layer": "Agent-Tool", "description": "Tool Access Protocol，AI Agent工具访问协议，标准化Agent对工具的发现、调用和管理", "status": "draft", "release_date": "2026", "source": "web_search_2026" }, { "id": "PROTO-048", "name": "AGP v2 (Agent Gateway Protocol v2)", "type": "agent_gateway_protocol", "layer": "Agent-Agent", "description": "Agent Gateway Protocol v2，Agent网关协议升级版，支持更安全的多Agent路由和通信", "status": "draft", "release_date": "2026", "source": "web_search_2026" }, { "id": "PROTO-049", "name": "MCP Extensions Protocol", "type": "agent_tool_protocol", "layer": "Agent-Tool", "description": "MCP扩展协议，2026-07-28 RC引入，支持MCP Server的扩展机制和插件系统", "status": "rc", "release_date": "2026-07", "source": "web_search_2026" }, { "id": "PROTO-050", "name": "MCP Stateless Core Protocol", "type": "agent_tool_protocol", "layer": "Agent-Tool", "description": "MCP无状态核心协议，2026-07-28 RC引入，将MCP核心改为无状态设计以支持水平扩展", "status": "rc", "release_date": "2026-07", "source": "web_search_2026" }, { "id": "PROTO-051", "name": "Agent Security Protocol (ASP)", "type": "agent_security_protocol", "layer": "Agent-Security", "description": "Agent安全协议，2026年MCP安全指南提出，解决AI工具隧道安全和Agent认证问题", "status": "proposal", "release_date": "2026", "source": "web_search_2026" }, { "id": "proto-agora-v2", "name": "AGORA Protocol v2", "type": "agent_communication", "description": "AGORA v2，牛津大学开发的自然语言动态协议，Agent自适应任务协商，支持复杂多Agent协作场景", "organization": "Oxford University", "key_features": [ "自然语言协议", "动态协议生成", "Agent自适应协商", "多Agent协作" ], "status": "research", "year": 2026, "sources": [ { "source_type": "web", "source_credibility": "A", "collected_at": "2026-06-03T13:52:48.985531Z" } ] }, { "id": "proto-apex-v2", "name": "APEX Standard v2", "type": "agent_interaction", "description": "APEX v2开放Agent交互协议标准，Registry + P2P架构，增强Agent发现和信任机制", "organization": "APEX Consortium", "key_features": [ "Registry架构", "P2P通信", "Agent发现", "信任机制" ], "status": "draft", "year": 2026, "sources": [ { "source_type": "web", "source_credibility": "B", "collected_at": "2026-06-03T13:52:48.985531Z" } ] }, { "id": "proto-agent-stack-2026", "name": "Agent Stack 2026 Protocol Suite", "type": "protocol_suite", "description": "2026年Agent协议栈：MCP + A2A + x402 + AP2 + Web Bot Auth + llms.txt，定义Agent生态完整协议层", "organization": "Industry Consortium", "key_features": [ "MCP工具层", "A2A通信层", "x402支付层", "Web Bot Auth认证层", "llms.txt发现层" ], "status": "production", "year": 2026, "sources": [ { "source_type": "web", "source_credibility": "A", "collected_at": "2026-06-03T13:52:48.985531Z" } ] }, { "id": "proto-mcp-stateless-rc", "name": "MCP Stateless Protocol RC", "type": "mcp_extension", "description": "MCP 2026-07-28 RC引入的无状态协议核心，支持StreamableHTTP传输和Tasks原语", "organization": "Anthropic", "key_features": [ "无状态核心", "StreamableHTTP", "Tasks原语", "MCP Apps" ], "status": "rc", "year": 2026, "sources": [ { "source_type": "web", "source_credibility": "A", "collected_at": "2026-06-03T13:52:48.985531Z" } ] }, { "id": "proto-a2a-v1-ibm-acp", "name": "A2A v1 + IBM ACP Merged", "type": "agent_communication", "description": "A2A v1.0正式版，IBM ACP已合并入A2A，150+组织采用，15000+ GitHub Stars", "organization": "Google/IBM", "key_features": [ "A2A正式版", "IBM ACP合并", "150+组织采用", "15000+ Stars" ], "status": "production", "year": 2026, "sources": [ { "source_type": "web", "source_credibility": "A", "collected_at": "2026-06-03T13:52:48.985531Z" } ] }, { "id": "PROTO-agui", "name": "AG-UI (Agent GUI Protocol)", "vendor": "CopilotKit", "version": "0.2", "layer": "Agent-UI", "description": "Agent GUI交互协议，定义Agent与前端UI组件的实时交互标准，SSE事件流驱动", "architecture": "Event Stream + Components", "transport": "SSE/WebSocket", "security": "CORS/认证", "adoption": "CopilotKit生态+扩展", "github_stars": "10000+", "status": "快速增长", "sources": [ { "source_type": "web", "source_credibility": "B", "collected_at": "2026-06-04T14:03:15.064Z" } ] }, { "id": "PROTO-x42", "name": "X42 (HTTP 402 Micro-Payment)", "vendor": "Coinbase/社区", "version": "1.0", "layer": "Agent-Payment", "description": "X42微支付协议，HTTP 402状态码驱动的机器对机器支付，与x402互补", "architecture": "HTTP-native", "transport": "HTTP/HTTPS", "security": "加密签名", "adoption": "增长中", "github_stars": "2000+", "status": "生产可用", "sources": [ { "source_type": "web", "source_credibility": "B", "collected_at": "2026-06-04T14:03:15.064Z" } ] }, { "id": "PROTO-mcp-server-cards", "name": "MCP Server Cards", "vendor": "Linux Foundation/MCP", "version": "draft", "layer": "Agent-Discovery", "description": "MCP Server Cards标准，通过.well-known URL暴露结构化Server元数据，支持浏览器和注册表发现", "architecture": "File-based Discovery", "transport": "HTTPS", "security": "TLS/签名", "adoption": "规划中", "github_stars": "N/A", "status": "规划", "sources": [ { "source_type": "web", "source_credibility": "B", "collected_at": "2026-06-04T14:03:15.064Z" } ] }, { "id": "PROTO-mpp", "name": "MPP (Merchant Payment Protocol)", "vendor": "行业联盟", "version": "draft", "layer": "Agent-Payment", "description": "商户支付协议，定义Agent与商户间的支付交互标准", "architecture": "Payment-native", "transport": "HTTPS", "security": "加密签名", "adoption": "增长中", "github_stars": "N/A", "status": "规划", "sources": [ { "source_type": "web", "source_credibility": "B", "collected_at": "2026-06-04T14:05:21.771Z" } ] }, { "id": "PROTO-agui", "name": "AGUI (Agent GUI Protocol)", "category": "agent-protocol", "description": "Agent GUI protocol standardizing how agents interact with user interfaces. One of the six key agent protocols in 2026.", "features": [ "agent-UI interaction", "GUI standardization", "user interaction protocol", "interface abstraction" ], "status": "active", "year": 2025, "sources": [ "https://www.mindstudio.ai/blog/six-agent-protocols-ai-builders-2026", "https://sanbi.ai/blog/agent-stack-protocols-2026" ] }, { "id": "PROTO-a2ui", "name": "A2UI (Agent to UI Protocol)", "category": "agent-protocol", "description": "Protocol for agent-to-user-interface communication. Complements A2A for inter-agent communication.", "features": [ "agent-to-UI", "interface communication", "user interaction" ], "status": "emerging", "year": 2025, "sources": [ "https://www.mindstudio.ai/blog/six-agent-protocols-ai-builders-2026" ] }, { "id": "PROTO-ap2", "name": "AP2 (Agent Protocol 2)", "category": "agent-protocol", "description": "Next-generation agent protocol. One of the six key protocols shaping AI agent interactions in 2026.", "features": [ "agent communication", "protocol standardization", "interoperability" ], "status": "emerging", "year": 2026, "sources": [ "https://www.mindstudio.ai/blog/six-agent-protocols-ai-builders-2026", "https://sanbi.ai/blog/agent-stack-protocols-2026" ] }, { "id": "PROTO-x42", "name": "X42 Protocol", "category": "payment-protocol", "description": "Payment protocol for AI agents enabling autonomous transactions. Part of the 2026 agent stack.", "features": [ "agent payments", "autonomous transactions", "economic protocol", "agent commerce" ], "status": "emerging", "year": 2026, "sources": [ "https://www.mindstudio.ai/blog/six-agent-protocols-ai-builders-2026", "https://sanbi.ai/blog/agent-stack-protocols-2026" ] }, { "id": "PROTO-streamable-http", "name": "StreamableHTTP (MCP Transport)", "category": "transport-protocol", "description": "New MCP transport layer introduced November 2025 replacing SSE. Uses chunked HTTP streaming for better reliability.", "features": [ "HTTP streaming", "SSE replacement", "chunked transfer", "MCP transport" ], "status": "active", "year": 2025, "sources": [ "https://www.speakeasy.com/mcp/release-notes", "https://workos.com/blog/everything-your-team-needs-to-know-about-mcp-in-2026" ] }, { "id": "PROTO-mcp-2026-rc", "name": "MCP 2026-07-28 Release Candidate", "type": "协议规范", "description": "Model Context Protocol 2026年7月28日发布候选版本：无状态协议核心 + Extensions框架，是MCP自发布以来最大架构升级", "key_features": [ "无状态协议核心", "Extensions扩展框架", "HTTP无状态调用", "大规模Agent部署支持", "向后兼容" ], "status": "release_candidate", "year": 2026, "maintainer": "Anthropic", "adoption": "9700万下载量(截至2026年)", "sources": [ { "source_type": "web", "source_credibility": "A", "article_url": "https://blog.modelcontextprotocol.io/posts/2026-07-28-release-candidate", "collected_at": "2026-06-16T22:31:38.000Z" }, { "source_type": "web", "source_credibility": "A", "article_url": "https://www.digitalapplied.com/blog/ai-agent-protocol-ecosystem-map-2026-mcp-a2a-acp-ucp", "collected_at": "2026-06-16T22:31:38.000Z" } ] }, { "id": "PROTO-a2a-2026", "name": "A2A (Agent-to-Agent) Protocol", "type": "Agent间通信协议", "description": "A2A协议是2026年AI Agent生态中Agent间通信的标准协议之一，与MCP(Agent-to-Tool)互补，解决多Agent协作问题", "key_features": [ "Agent间通信", "任务委派", "状态同步", "多Agent协作" ], "status": "emerging", "year": 2026, "maintainer": "社区", "adoption": "快速增长", "sources": [ { "source_type": "web", "source_credibility": "B", "article_url": "https://www.digitalapplied.com/blog/ai-agent-protocol-ecosystem-map-2026-mcp-a2a-acp-ucp", "collected_at": "2026-06-16T22:31:38.000Z" } ] }, { "name": "Agent2Agent Protocol (A2A)", "type": "Communication Protocol", "description": "An open protocol enabling AI agents to securely exchange information, discover capabilities, and coordinate actions across different enterprise systems.", "status": "Active", "features": [ "Secure information exchange", "Agent capability discovery", "Cross-agent action coordination", "Enterprise interoperability" ], "sources": [ "https://developers.googleblog.com/en/a2a-a-new-era-of-agent-interoperability", "https://arxiv.org/html/2505.02279v1", "https://www.linkedin.com/posts/rakeshgohel01_starting-2025-ai-agent-interoperability-activity-7330572530061103105-ssMh", "https://www.digitalapplied.com/blog/ai-agent-protocol-ecosystem-map-2026-mcp-a2a-acp-ucp", "https://blogs.cisco.com/ai/mcp-and-a2a-a-network-engineers-mental-model-for-agentic-ai", "https://aws.amazon.com/blogs/opensource/open-protocols-for-agent-interoperability-part-1-inter-agent-communication-on-mcp", "https://www.reddit.com/r/datascience/comments/1lq79vo/a_breakdown_of_a2a_mcp_and_agentic" ], "id": "PROTO-067" }, { "name": "Model Context Protocol (MCP)", "type": "Tool Exposure Protocol", "description": "A standardized protocol designed to expose AI tools and models to agents, enabling seamless integration and interaction between AI systems and external resources.", "status": "Active", "features": [ "Tool exposure to AI agents", "Seamless system integration", "External resource interaction", "Complementary to A2A" ], "sources": [ "https://arxiv.org/html/2505.02279v1", "https://www.digitalapplied.com/blog/ai-agent-protocol-ecosystem-map-2026-mcp-a2a-acp-ucp", "https://www.youtube.com/watch?v=BMDFPOyezH4", "https://blogs.cisco.com/ai/mcp-and-a2a-a-network-engineers-mental-model-for-agentic-ai", "https://aws.amazon.com/blogs/opensource/open-protocols-for-agent-interoperability-part-1-inter-agent-communication-on-mcp", "https://www.reddit.com/r/datascience/comments/1lq79vo/a_breakdown_of_a2a_mcp_and_agentic" ], "id": "PROTO-068" }, { "name": "Agent Communication Protocol (ACP)", "type": "Interoperability Standard", "description": "An open governance standard allowing AI agents to communicate across diverse frameworks and technologies, promoting multi-framework agent interoperability.", "status": "Active", "features": [ "Multi-framework compatibility", "Open governance model", "Cross-technology communication", "Standardized agent interaction" ], "sources": [ "https://arxiv.org/html/2505.02279v1", "https://www.digitalapplied.com/blog/ai-agent-protocol-ecosystem-map-2026-mcp-a2a-acp-ucp", "https://www.ibm.com/think/topics/agent-communication-protocol" ], "id": "PROTO-069" }, { "name": "Linux Foundation Agent Communication Protocol (ACP)", "type": "Open Standard", "description": "An open standard developed under the Linux Foundation with BeeAI as its reference implementation, emphasizing transparent community-driven governance for agent interoperability.", "status": "Active", "features": [ "Community-driven governance", "Transparent development", "BeeAI reference implementation", "Linux Foundation backing" ], "sources": [ "https://agentcommunicationprotocol.dev/introduction/welcome" ], "id": "PROTO-070" }, { "name": "BeeAI", "type": "Reference Implementation", "description": "The reference implementation for the Linux Foundation's Agent Communication Protocol (ACP), serving as a practical example of how agents can interoperate using standardized protocols.", "status": "Active", "features": [ "ACP reference implementation", "Demonstrates protocol interoperability", "Linux Foundation project", "Practical agent communication example" ], "sources": [ "https://agentcommunicationprotocol.dev/introduction/welcome" ], "id": "PROTO-071" }, { "name": "OpenTelemetry (OTEL)", "type": "Observability Framework", "description": "An observability framework that will be supported alongside A2A and MCP to enable monitoring and traceability for agentic AI systems.", "status": "Upcoming Support", "features": [ "System observability", "Tracing and monitoring", "Integration with A2A/MCP", "Open standard" ], "sources": [ "https://aws.amazon.com/blogs/opensource/open-protocols-for-agent-interoperability-part-1-inter-agent-communication-on-mcp" ], "id": "PROTO-072" } ], "memory_systems": [ { "id": "MEM-001", "name": "mem0", "type": "向量+图谱", "architecture": "混合", "features": [ "自动记忆提取", "用户画像", "时间衰减" ], "open_source": true, "license": "MIT", "maturity": "生产可用", "pricing": "免费/Pro$29/月", "language": "Python" }, { "id": "MEM-002", "name": "Zep", "type": "向量+图谱", "architecture": "混合", "features": [ "对话记忆", "实体提取", "知识图谱" ], "open_source": true, "license": "Apache-2.0", "maturity": "生产可用", "pricing": "免费/Cloud付费", "language": "Python/TS" }, { "id": "MEM-003", "name": "Letta (MemGPT)", "type": "分层记忆", "architecture": "上下文+归档", "features": [ "虚拟上下文管理", "自动分页", "无限上下文" ], "open_source": true, "license": "Apache-2.0", "maturity": "生产可用", "pricing": "免费/Cloud付费", "language": "Python" }, { "id": "MEM-004", "name": "Cognee", "type": "知识图谱", "architecture": "Graph RAG", "features": [ "自动知识图谱构建", "实体关系推理" ], "open_source": true, "license": "Apache-2.0", "maturity": "早期", "pricing": "免费", "language": "Python" }, { "id": "MEM-005", "name": "Graphiti", "type": "知识图谱", "architecture": "时序图谱", "features": [ "自动演化图谱", "时间推理", "非结构化输入" ], "open_source": true, "license": "Apache-2.0", "maturity": "早期", "pricing": "免费", "language": "Python" }, { "id": "MEM-006", "name": "Hindsight", "type": "对话记忆", "architecture": "向量", "features": [ "对话摘要", "关键信息提取", "上下文窗口" ], "open_source": true, "license": "MIT", "maturity": "早期", "pricing": "免费", "language": "Python" }, { "id": "MEM-007", "name": "LangMem", "type": "工作流记忆", "architecture": "LangGraph集成", "features": [ "LangGraph原生记忆", "工作流编排集成", "记忆工具链", "语义搜索" ], "open_source": true, "license": "MIT", "maturity": "生产可用", "pricing": "免费", "language": "Python", "benchmarks": "LangGraph生态", "sources": [ { "source_type": "web", "source_credibility": "B", "collected_at": "2026-05-30T13:38:54.619Z" } ] }, { "id": "MEM-008", "name": "Memvid", "type": "多模态记忆", "architecture": "视频编码", "features": [ "视频记忆编码", "多模态记忆检索", "视觉上下文保存" ], "open_source": true, "license": "MIT", "maturity": "实验性", "pricing": "免费", "language": "Python", "sources": [ { "source_type": "web", "source_credibility": "C", "collected_at": "2026-05-30T13:38:54.619Z" } ] }, { "id": "MEM-009", "name": "MemoClaw", "type": "跨平台记忆", "architecture": "分布式", "features": [ "Agent记忆管理", "跨平台记忆同步", "结构化记忆存储" ], "open_source": true, "license": "MIT", "maturity": "生产可用", "pricing": "免费/付费", "language": "Python/TS", "sources": [ { "source_type": "web", "source_credibility": "C", "collected_at": "2026-05-30T13:38:54.619Z" } ] }, { "id": "MEM-010", "name": "EverMind", "type": "长期记忆", "architecture": "个性化", "features": [ "Zep替代方案", "长期记忆管理", "对话理解", "个性化记忆" ], "open_source": true, "license": "MIT", "maturity": "生产可用", "pricing": "免费/付费", "language": "Python", "sources": [ { "source_type": "web", "source_credibility": "C", "collected_at": "2026-05-30T13:38:54.619Z" } ] }, { "id": "MEM-011", "name": "Pinecone Memory", "type": "向量记忆", "architecture": "托管向量", "features": [ "Pinecone托管向量", "Serverless部署", "命名空间隔离", "高可用" ], "open_source": false, "license": "Commercial", "maturity": "生产可用", "pricing": "按用量付费", "language": "Cloud API", "sources": [ { "source_type": "web", "source_credibility": "A", "collected_at": "2026-05-30T13:38:54.619Z" } ] }, { "id": "MEM-012", "name": "Mem0 v2 (Graph Memory)", "type": "向量+图谱", "architecture": "图记忆增强", "features": [ "图记忆功能", "自动记忆提取", "跨会话持久化", "55K+ Stars", "LoCoMo/LongMemEval/BEAM基准测试领先" ], "open_source": true, "license": "MIT", "maturity": "生产可用", "pricing": "免费/Pro$29/月/图记忆$249/月", "language": "Python/TS", "benchmarks": "LoCoMo/LongMemEval/BEAM标准", "sources": [ { "source_type": "web", "source_credibility": "A", "collected_at": "2026-05-30T13:38:54.619Z" } ] }, { "id": "MEM-013", "name": "Hindsight v2", "type": "memory_system", "description": "Hindsight v2，2026年Agent记忆系统，准确率91.4%，支持层次化记忆架构和知识图谱", "accuracy": "91.4%", "release_date": "2026", "source": "web_search_2026" }, { "id": "MEM-014", "name": "Cognee v2", "type": "memory_system", "description": "Cognee v2，2026年知识图谱记忆系统升级版，支持多Agent共享记忆和类型化关系", "release_date": "2026", "source": "web_search_2026" }, { "id": "MEM-015", "name": "Zep v2", "type": "memory_system", "description": "Zep v2，2026年Agent长期记忆系统升级版，支持结构化知识提取和对话记忆管理", "release_date": "2026", "source": "web_search_2026" }, { "id": "MEM-016", "name": "LangChain Memory", "category": "memory-framework", "description": "LangChain's built-in memory module for AI agents. Supports conversation buffer, summary, and entity memory patterns.", "features": [ "conversation buffer", "summary memory", "entity memory", "LangChain integration" ], "status": "active", "year": 2024, "sources": [ "https://machinelearningmastery.com/the-6-best-ai-agent-memory-frameworks-you-should-try-in-2026" ] }, { "id": "MEM-017", "name": "LlamaIndex Memory", "category": "memory-framework", "description": "LlamaIndex's memory module for AI agents with document-indexed retrieval and context management.", "features": [ "document indexing", "retrieval memory", "context management", "LlamaIndex integration" ], "status": "active", "year": 2024, "sources": [ "https://machinelearningmastery.com/the-6-best-ai-agent-memory-frameworks-you-should-try-in-2026" ] }, { "id": "MEM-018", "name": "ICLR 2026 Agent Memory Workshop", "category": "research", "description": "ICLR 2026 Workshop on Memory for LLM-Based Agentic Systems. Advances design of memory layer for agentic systems.", "features": [ "academic research", "memory architecture", "agentic systems", "workshop proceedings" ], "status": "active", "year": 2026, "sources": [ "https://iclr.cc/virtual/2026/workshop/10000792" ] }, { "id": "MEM-019", "name": "Mem0 State of Agent Memory 2026", "category": "benchmark", "description": "Comprehensive report on AI agent memory in 2026 by Mem0. Production engineering discipline with real benchmarks and measurable trade-offs.", "features": [ "benchmark report", "production metrics", "memory evaluation", "industry survey" ], "status": "active", "year": 2026, "sources": [ "https://mem0.ai/blog/state-of-ai-agent-memory-2026" ] }, { "id": "MEM-020", "name": "SitePoint Agent Memory Guide", "category": "guide", "description": "Complete guide to agent memory in 2026 covering memory architecture concepts and five most destructive memory failure modes.", "features": [ "memory architecture", "failure modes", "production guide", "best practices" ], "status": "active", "year": 2026, "sources": [ "https://www.sitepoint.com/the-new-reality-of-agent-memory-the-complete-guide-2026" ] }, { "id": "MEM-021", "name": "LoCoMo Benchmark", "full_name": "LoCoMo Long-term Conversation Memory Benchmark", "category": "记忆基准", "description": "LoCoMo是2026年Agent记忆架构比较的标准基准之一，用于评估长期对话记忆能力，与LongMemEval和BEAM并列为三大记忆基准", "key_features": [ "长期对话记忆评估", "多轮交互测试", "记忆检索准确率", "行业标准基准" ], "status": "active", "year": 2026, "maintainer": "学术界", "sources": [ { "source_type": "web", "source_credibility": "A", "article_url": "https://mem0.ai/blog/state-of-ai-agent-memory-2026", "collected_at": "2026-06-16T22:31:38.000Z" } ] }, { "id": "MEM-022", "name": "LongMemEval Benchmark", "full_name": "LongMemEval Long-term Memory Retrieval Benchmark", "category": "记忆基准", "description": "LongMemEval (ICLR 2025)是长期记忆检索评估基准，每个问题约48个会话(S变体500题、~115K tokens)，已成为Agent记忆系统标准评估工具", "key_features": [ "48会话/问题", "500题(S变体)", "~115K tokens/问题", "ICLR 2025论文" ], "status": "active", "year": 2025, "maintainer": "学术界", "sources": [ { "source_type": "web", "source_credibility": "A", "article_url": "https://github.com/rohitg00/agentmemory/blob/main/benchmark/COMPARISON.md", "collected_at": "2026-06-16T22:31:38.000Z" }, { "source_type": "web", "source_credibility": "A", "article_url": "https://mem0.ai/blog/state-of-ai-agent-memory-2026", "collected_at": "2026-06-16T22:31:38.000Z" } ] }, { "id": "MEM-023", "name": "BEAM Benchmark", "full_name": "BEAM Agent Memory Benchmark", "category": "记忆基准", "description": "BEAM是2026年三大Agent记忆基准之一(与LoCoMo和LongMemEval并列)，用于全面评估Agent记忆架构性能", "key_features": [ "Agent记忆综合评估", "与LoCoMo/LongMemEval并列", "行业标准" ], "status": "active", "year": 2026, "maintainer": "学术界", "sources": [ { "source_type": "web", "source_credibility": "A", "article_url": "https://mem0.ai/blog/state-of-ai-agent-memory-2026", "collected_at": "2026-06-16T22:31:38.000Z" } ] }, { "name": "Letta", "type": "Framework", "description": "An open-source community-led approach to AI agent memory management.", "features": [ "Open source", "Community-driven", "Long-term memory" ], "status": "Battle-Tested", "sources": [ "https://medium.com/asymptotic-spaghetti-integration/from-beta-to-battle-tested-picking-between-letta-mem0-zep-for-ai-memory-6850ca8703d1", "https://forum.letta.com/t/agent-memory-letta-vs-mem0-vs-zep-vs-cognee/88" ], "id": "MEM-024" }, { "name": "IBM AI Agent Memory", "type": "Concept/Resource", "description": "IBM's perspective on AI agent memory, focusing on Long-term memory (LTM) for personalization and intelligence.", "features": [ "Long-term memory (LTM)", "Personalization", "Cross-session recall" ], "status": "Reference", "sources": [ "https://www.ibm.com/think/topics/ai-agent-memory" ], "id": "MEM-025" } ], "model_apis": [ { "id": "API-001", "name": "OpenAI API", "vendor": "OpenAI", "type": "LLM API", "models": [ "GPT-4o", "GPT-4o-mini", "o1", "o3-mini", "DALL-E 3", "Whisper", "TTS" ], "pricing_input": "$2.5/1M tokens (GPT-4o)", "pricing_output": "$10/1M tokens (GPT-4o)", "context_window": "128K", "features": [ "Chat Completions", "Function Calling", "Vision", "Structured Output", "Batch API", "Fine-tuning", "Assistants API" ], "rate_limit": "Tier1: 500 RPM", "sdk_languages": [ "Python", "Node.js", "Go", "Java", "C#" ], "docs": "https://platform.openai.com/docs", "status": "生产可用" }, { "id": "API-002", "name": "Anthropic API", "vendor": "Anthropic", "type": "LLM API", "models": [ "Claude 3.5 Sonnet", "Claude 3.5 Haiku", "Claude 3 Opus" ], "pricing_input": "$3/1M tokens (Sonnet)", "pricing_output": "$15/1M tokens (Sonnet)", "context_window": "200K", "features": [ "Messages API", "Tool Use", "Vision", "MCP Support", "Prompt Caching", "Extended Thinking" ], "rate_limit": "Tier1: 1000 RPM", "sdk_languages": [ "Python", "Node.js" ], "docs": "https://docs.anthropic.com", "status": "生产可用" }, { "id": "API-003", "name": "Google Gemini API", "vendor": "Google", "type": "LLM API", "models": [ "Gemini 2.0 Flash", "Gemini 2.5 Pro", "Gemini 1.5 Pro" ], "pricing_input": "$1.25/1M tokens (Flash)", "pricing_output": "$5/1M tokens (Flash)", "context_window": "1M-2M", "features": [ "Generate Content", "Function Calling", "Vision", "Code Execution", "Grounding", "Context Caching" ], "rate_limit": "Free: 15 RPM, Pay: 2000 RPM", "sdk_languages": [ "Python", "Node.js", "Go", "Java", "Dart" ], "docs": "https://ai.google.dev/docs", "status": "生产可用" }, { "id": "API-004", "name": "DeepSeek API", "vendor": "DeepSeek", "type": "LLM API", "models": [ "DeepSeek-V3", "DeepSeek-R1" ], "pricing_input": "$0.27/1M tokens", "pricing_output": "$1.1/1M tokens", "context_window": "128K", "features": [ "Chat Completions", "FIM", "JSON Output", "Function Calling" ], "rate_limit": "按余额", "sdk_languages": [ "OpenAI兼容" ], "docs": "https://platform.deepseek.com/api-docs", "status": "生产可用" }, { "id": "API-005", "name": "智谱AI API", "vendor": "智谱AI", "type": "LLM API", "models": [ "GLM-4", "GLM-4V", "GLM-4-Flash" ], "pricing_input": "¥0.1/1M tokens (Flash)", "pricing_output": "¥0.1/1M tokens (Flash)", "context_window": "128K", "features": [ "Chat Completions", "Function Calling", "Vision", "Web Search", "Code Interpreter" ], "rate_limit": "按套餐", "sdk_languages": [ "Python", "Node.js", "Java" ], "docs": "https://open.bigmodel.cn/dev/api", "status": "生产可用" }, { "id": "API-006", "name": "Mistral API", "vendor": "Mistral AI", "type": "LLM API", "models": [ "Mistral Large", "Mistral Medium", "Codestral", "Pixtral" ], "pricing_input": "$2/1M tokens (Large)", "pricing_output": "$6/1M tokens (Large)", "context_window": "128K", "features": [ "Chat Completions", "Function Calling", "Vision", "Embeddings", "Agents API" ], "rate_limit": "按套餐", "sdk_languages": [ "Python", "Node.js" ], "docs": "https://docs.mistral.ai", "status": "生产可用" }, { "id": "API-007", "name": "Cohere API", "vendor": "Cohere", "type": "LLM/RAG API", "models": [ "Command R+", "Command R", "Embed v3", "Rerank v3" ], "pricing_input": "$2.5/1M tokens (R+)", "pricing_output": "$10/1M tokens (R+)", "context_window": "128K", "features": [ "Chat", "RAG", "Embeddings", "Rerank", "Connectors", "Fine-tuning" ], "rate_limit": "按套餐", "sdk_languages": [ "Python", "Node.js" ], "docs": "https://docs.cohere.com", "status": "生产可用" }, { "id": "API-008", "name": "Groq API", "vendor": "Groq", "type": "推理加速API", "models": [ "Llama 3.3 70B", "Mixtral 8x7B", "Gemma 2 9B" ], "pricing_input": "$0.59/1M tokens", "pricing_output": "$0.79/1M tokens", "context_window": "128K", "features": [ "超低延迟推理", "OpenAI兼容", "Chat Completions", "Vision" ], "rate_limit": "按套餐", "sdk_languages": [ "OpenAI兼容" ], "docs": "https://console.groq.com/docs", "status": "生产可用" }, { "id": "API-009", "name": "Together AI API", "vendor": "Together AI", "type": "推理平台", "models": [ "200+开源模型", "Llama 3", "Qwen 2.5", "DeepSeek" ], "pricing_input": "$0.18/1M tokens (Llama3)", "pricing_output": "$0.18/1M tokens", "context_window": "128K", "features": [ "Serverless推理", "Fine-tuning", "Embeddings", "Image Gen" ], "rate_limit": "按余额", "sdk_languages": [ "OpenAI兼容" ], "docs": "https://docs.together.ai", "status": "生产可用" }, { "id": "API-010", "name": "Fireworks AI API", "vendor": "Fireworks AI", "type": "推理平台", "models": [ "Llama 3.3", "Qwen 2.5", "Mixtral", "Whisper" ], "pricing_input": "$0.2/1M tokens", "pricing_output": "$0.2/1M tokens", "context_window": "128K", "features": [ "Serverless推理", "Fine-tuning", "Vision", "Function Calling" ], "rate_limit": "按余额", "sdk_languages": [ "OpenAI兼容" ], "docs": "https://docs.fireworks.ai", "status": "生产可用" }, { "id": "API-011", "name": "Perplexity API", "vendor": "Perplexity", "type": "搜索AI API", "models": [ "Sonar", "Sonar Pro", "Sonar Reasoning" ], "pricing_input": "$1/1M tokens", "pricing_output": "$1/1M tokens", "context_window": "128K", "features": [ "实时搜索", "引用来源", "Pro深度搜索" ], "rate_limit": "按套餐", "sdk_languages": [ "Python", "Node.js" ], "docs": "https://docs.perplexity.ai", "status": "生产可用" }, { "id": "API-012", "name": "ElevenLabs API", "vendor": "ElevenLabs", "type": "TTS/STT API", "models": [ "eleven_multilingual_v2", "eleven_turbo_v2", "eleven_monolingual_v1" ], "pricing_input": "$0.18/1K chars", "pricing_output": "N/A", "context_window": "N/A", "features": [ "TTS", "STT", "Voice Cloning", "Sound Effects", "Dubbing" ], "rate_limit": "按套餐", "sdk_languages": [ "Python", "Node.js" ], "docs": "https://elevenlabs.io/docs/api-reference", "status": "生产可用" }, { "id": "API-013", "name": "Claude Agent SDK API", "full_name": "Anthropic Agent SDK API", "category": "Agent API", "description": "Anthropic Claude Agent SDK于2026年6月15日开始从独立月度Agent SDK额度计费，支持文件存储和事件构建", "maintainer": "Anthropic", "status": "active", "features": [ "独立计费", "文件存储", "事件构建", "Agent原生" ] }, { "id": "API-014", "name": "Google ADK API", "full_name": "Google Agent Development Kit API", "category": "Agent API", "description": "Google ADK提供Agent开发API，优化Google Cloud集成，支持层级Agent树和多Agent编排", "maintainer": "Google", "status": "active", "features": [ "Google Cloud优化", "层级Agent树", "多Agent编排", "Vertex AI集成" ] }, { "id": "API-015", "name": "OpenAI Agents API", "full_name": "OpenAI Agents API", "category": "Agent API", "description": "OpenAI Agents SDK配套API，支持Handoff模式、工具调用和多Agent协作，1030万月下载量", "maintainer": "OpenAI", "status": "active", "features": [ "Handoff模式", "工具调用", "多Agent协作", "大规模采用" ] }, { "id": "API-016", "name": "MCP Tasks API", "full_name": "MCP Tasks API", "category": "协议API", "description": "MCP 2026规范引入的Tasks API，支持异步任务调度和长运行操作管理", "maintainer": "Anthropic", "status": "active", "features": [ "异步任务", "长运行操作", "任务状态追踪", "2026规范" ] } ], "vector_dbs": [ { "id": "VDB-001", "name": "Pinecone", "type": "托管向量数据库", "pricing": "免费/Starter$25/月", "features": [ "自动扩缩", "元数据过滤", "命名空间", "稀疏向量" ], "open_source": false, "maturity": "生产可用", "language": "Python/Node.js" }, { "id": "VDB-002", "name": "Weaviate", "type": "向量+图谱数据库", "pricing": "免费/Cloud付费", "features": [ "混合搜索", "多模态", "RAG模块", "GraphQL" ], "open_source": true, "license": "BSD-3", "maturity": "生产可用", "language": "Python/Node.js/Go/Java" }, { "id": "VDB-003", "name": "Qdrant", "type": "向量数据库", "pricing": "免费/Cloud付费", "features": [ "高性嫩过滤", "量化压缩", "命名空间", "Payload过滤" ], "open_source": true, "license": "Apache-2.0", "maturity": "生产可用", "language": "Python/Node.js/Rust" }, { "id": "VDB-004", "name": "ChromaDB", "type": "嵌入式向量数据库", "pricing": "免费", "features": [ "嵌入式/Server", "自动嵌入", "元数据", "相似度搜索" ], "open_source": true, "license": "Apache-2.0", "maturity": "生产可用", "language": "Python/JS" }, { "id": "VDB-005", "name": "Milvus", "type": "分布式向量数据库", "pricing": "免费/Cloud付费", "features": [ "十亿级向量", "GPU加速", "多索引", "标量过滤" ], "open_source": true, "license": "Apache-2.0", "maturity": "生产可用", "language": "Python/Node.js/Go/Java/C++" }, { "id": "VDB-006", "name": "pgvector", "type": "PostgreSQL扩展", "pricing": "免费", "features": [ "SQL集成", "IVFFlat/HNSW索引", "无需新基础设施" ], "open_source": true, "license": "PostgreSQL", "maturity": "生产可用", "language": "SQL" }, { "id": "VDB-007", "name": "LanceDB", "type": "Serverless向量数据库", "pricing": "免费", "features": [ "嵌入式", "列存储", "版本管理", "零拷贝" ], "open_source": true, "license": "Apache-2.0", "maturity": "早期", "language": "Python/Node.js/Rust" }, { "id": "VDB-008", "name": "Neo4j + Vector", "type": "图+向量数据库", "pricing": "免费/Enterprise", "features": [ "知识图谱+向量", "Cypher查询", "GDS算法" ], "open_source": true, "license": "GPLv3(社区)", "maturity": "生产可用", "language": "Python/JS/Java/Go" }, { "id": "VDB-pinecone-v2", "name": "Pinecone Serverless v2", "type": "managed", "company": "Pinecone", "language": "Cloud API", "license": "Commercial", "key_features": [ "Serverless架构", "数十亿向量存储", "新定价模型", "命名空间隔离" ], "status": "production", "year": 2026, "sources": [ { "source_type": "web", "source_credibility": "A", "collected_at": "2026-05-30T13:38:30.808Z" } ] }, { "id": "VDB-chroma-cloud", "name": "Chroma Cloud", "type": "managed", "company": "Chroma", "language": "Python/Cloud API", "license": "Commercial", "key_features": [ "Chroma Cloud分布式模式", "2026年初发布", "嵌入式优先", "开发者友好" ], "status": "production", "year": 2026, "sources": [ { "source_type": "web", "source_credibility": "B", "collected_at": "2026-05-30T13:38:30.808Z" } ] }, { "id": "VDB-qdrant-hybrid", "name": "Qdrant Hybrid Search v2", "type": "open_source", "company": "Qdrant", "language": "Rust", "license": "Apache-2.0", "key_features": [ "混合搜索升级", "稀疏+密集向量", "Rust高性能", "过滤优化" ], "status": "production", "year": 2026, "sources": [ { "source_type": "web", "source_credibility": "B", "collected_at": "2026-05-30T13:38:30.808Z" } ] }, { "id": "VDB-weaviate-v2", "name": "Weaviate v2", "type": "open_source", "company": "Weaviate", "language": "Go", "license": "BSD-3", "key_features": [ "多模态搜索", "生成式搜索", "模块化架构", "GraphQL API" ], "status": "production", "year": 2026, "sources": [ { "source_type": "web", "source_credibility": "B", "collected_at": "2026-05-30T13:38:30.808Z" } ] }, { "id": "VDB-milvus-v3", "name": "Milvus v3", "type": "open_source", "company": "Zilliz", "language": "Go/C++", "license": "Apache-2.0", "key_features": [ "云原生架构", "十亿级向量", "GPU加速", "多索引支持" ], "status": "production", "year": 2026, "sources": [ { "source_type": "web", "source_credibility": "B", "collected_at": "2026-05-30T13:38:30.808Z" } ] }, { "id": "VDB-pgvector-v2", "name": "pgvector v2 (PostgreSQL)", "type": "extension", "company": "社区", "language": "C", "license": "PostgreSQL", "key_features": [ "PostgreSQL扩展", "HNSW索引", "IVFFlat索引", "半精度向量" ], "status": "production", "year": 2025, "sources": [ { "source_type": "web", "source_credibility": "A", "collected_at": "2026-05-30T13:38:30.808Z" } ] }, { "id": "VDB-vald", "name": "Vald", "type": "open_source", "company": "Yahoo Japan", "language": "Go", "license": "Apache-2.0", "key_features": [ "云原生", "自动索引", "高可用", "gRPC接口" ], "status": "production", "year": 2025, "sources": [ { "source_type": "web", "source_credibility": "B", "collected_at": "2026-05-30T13:38:30.808Z" } ] }, { "id": "VDB-vespa", "name": "Vespa", "type": "open_source", "company": "Yahoo/Vespa.ai", "language": "Java/C++", "license": "Apache-2.0", "key_features": [ "实时向量+文本搜索", "低延迟", "自动缩放", "内容推荐" ], "status": "production", "year": 2025, "sources": [ { "source_type": "web", "source_credibility": "B", "collected_at": "2026-05-30T13:38:30.808Z" } ] }, { "id": "VDB-typesense", "name": "Typesense", "type": "open_source", "company": "Typesense", "language": "C++", "license": "GPL-3.0", "key_features": [ "即时搜索", "向量搜索集成", "拼写纠错", "嵌入式" ], "status": "production", "year": 2025, "sources": [ { "source_type": "web", "source_credibility": "B", "collected_at": "2026-05-30T13:38:30.808Z" } ] }, { "id": "VDB-marqo", "name": "Marqo", "type": "open_source", "company": "Marqo", "language": "Python", "license": "Apache-2.0", "key_features": [ "端到端向量搜索", "内置嵌入模型", "多模态", "张量搜索" ], "status": "production", "year": 2025, "sources": [ { "source_type": "web", "source_credibility": "B", "collected_at": "2026-05-30T13:38:30.808Z" } ] }, { "id": "VDB-supabase-pgvector", "name": "Supabase pgvector", "type": "managed", "company": "Supabase", "language": "SQL/TypeScript", "license": "PostgreSQL/Apache", "key_features": [ "Supabase托管pgvector", "50亿美元估值2026", "PostgreSQL生态", "实时订阅", "Edge Functions集成" ], "status": "production", "year": 2026, "sources": [ { "source_type": "web", "source_credibility": "A", "collected_at": "2026-06-01T13:48:20.198Z" } ] }, { "id": "VDB-faiss-v2", "name": "FAISS v2 (Meta)", "type": "open_source", "company": "Meta", "language": "C++/Python", "license": "MIT", "key_features": [ "Meta开源向量搜索库", "GPU加速", "十亿级索引", "量化压缩", "2026性能优化" ], "status": "production", "year": 2026, "sources": [ { "source_type": "web", "source_credibility": "A", "collected_at": "2026-06-01T13:48:20.198Z" } ] }, { "id": "VDB-lancedb-v2", "name": "LanceDB v2", "type": "open_source", "company": "LanceDB", "language": "Rust/Python/TS", "license": "Apache-2.0", "key_features": [ "Serverless向量数据库v2", "Lance列式格式", "多模态支持", "零拷贝读取", "版本管理" ], "status": "production", "year": 2026, "sources": [ { "source_type": "web", "source_credibility": "B", "collected_at": "2026-06-01T13:48:20.198Z" } ] }, { "id": "VDB-chromadb-cloud-v2", "name": "Chroma Cloud v2", "type": "managed", "company": "Chroma", "language": "Python/Rust", "license": "Apache-2.0", "key_features": [ "Chroma Cloud分布式v2", "嵌入式优先", "自动嵌入", "混合搜索", "2026全面发布" ], "status": "production", "year": 2026, "sources": [ { "source_type": "web", "source_credibility": "B", "collected_at": "2026-06-01T13:48:20.198Z" } ] }, { "id": "VDB-mongodb-atlas", "name": "MongoDB Atlas Vector Search", "type": "vector_database", "description": "MongoDB Atlas向量搜索，2026年主流向量数据库之一，支持原生向量索引和混合查询", "vendor": "MongoDB", "deployment": "cloud", "release_date": "2026", "source": "web_search_2026" }, { "id": "VDB-milvus-v4", "name": "Milvus v4", "type": "vector_database", "description": "Milvus v4，2026年最新版本，增强分布式性能和GPU加速向量搜索", "vendor": "Zilliz", "deployment": "hybrid", "release_date": "2026", "source": "web_search_2026" }, { "id": "VDB-qdrant-v3", "name": "Qdrant v3", "type": "vector_database", "description": "Qdrant v3，2026年最新版本，改进混合搜索性能和过滤能力", "vendor": "Qdrant", "deployment": "hybrid", "release_date": "2026", "source": "web_search_2026" }, { "id": "VDB-weaviate-v3", "name": "Weaviate v3", "type": "vector_database", "description": "Weaviate v3，2026年最新版本，增强多模态搜索和Agent记忆集成", "vendor": "Weaviate", "deployment": "hybrid", "release_date": "2026", "source": "web_search_2026" }, { "id": "VDB-pinecone-v3", "name": "Pinecone v3", "type": "vector_database", "description": "Pinecone v3，2026年最新版本，改进Serverless性能和命名空间隔离", "vendor": "Pinecone", "deployment": "cloud", "release_date": "2026", "source": "web_search_2026" }, { "id": "VDB-turbopuffer", "name": "Turbopuffer", "type": "serverless_vector_search", "description": "Turbopuffer Serverless向量搜索，2026年新进入者，低延迟托管向量搜索服务", "company": "Turbopuffer", "language": "Cloud API", "license": "Commercial", "key_features": [ "Serverless", "低延迟", "托管向量搜索", "自动扩缩" ], "status": "production", "year": 2026, "sources": [ { "source_type": "web", "source_credibility": "B", "collected_at": "2026-06-03T13:52:48.985531Z" } ] }, { "id": "VDB-myscale", "name": "MyScale", "type": "sql_vector_db", "description": "MyScale SQL向量数据库，基于ClickHouse的向量搜索，支持SQL+向量混合查询", "company": "MyScale", "language": "SQL/Python", "license": "Apache-2.0", "key_features": [ "SQL+向量混合", "ClickHouse引擎", "高性能分析", "结构化+非结构化" ], "status": "production", "year": 2026, "sources": [ { "source_type": "web", "source_credibility": "B", "collected_at": "2026-06-03T13:52:48.985531Z" } ] }, { "id": "VDB-singlestore", "name": "SingleStore Vector", "type": "distributed_sql_vector", "description": "SingleStore分布式SQL+向量数据库，支持实时分析和向量搜索", "company": "SingleStore", "language": "SQL", "license": "Commercial", "key_features": [ "分布式SQL", "实时分析", "向量搜索", "混合事务分析" ], "status": "production", "year": 2026, "sources": [ { "source_type": "web", "source_credibility": "B", "collected_at": "2026-06-03T13:52:48.985531Z" } ] }, { "id": "VDB-kdbai", "name": "KDB.AI", "type": "timeseries_vector_db", "description": "KDB.AI时序向量数据库，专为金融和IoT时序数据+向量搜索设计", "company": "KX Systems", "language": "Python/q", "license": "Commercial", "key_features": [ "时序数据", "金融级", "向量搜索", "实时分析" ], "status": "production", "year": 2026, "sources": [ { "source_type": "web", "source_credibility": "B", "collected_at": "2026-06-03T13:52:48.985531Z" } ] }, { "id": "VDB-objectbox", "name": "ObjectBox Vector Database", "category": "on-device-vector-db", "description": "On-device vector database engineered for phones, ECUs, and other restricted environments. The missing on-device memory layer for AI.", "features": [ "on-device", "edge computing", "mobile support", "embedded systems", "low resource" ], "status": "active", "year": 2026, "sources": [ "https://objectbox.io/262454-2" ] }, { "id": "VDB-cassandra5", "name": "Apache Cassandra 5.0 (Vector Search)", "category": "distributed-vector-db", "description": "Apache Cassandra 5.0 introduced vector search support for AI and machine learning workloads.", "features": [ "vector search", "distributed", "cloud-native", "scalable", "AI workloads" ], "status": "active", "year": 2025, "sources": [ "https://www.instaclustr.com/education/vector-database/best-open-source-vector-database-solutions-top-5-in-2026" ] }, { "id": "VDB-pinecone-2026", "name": "Pinecone 2026", "full_name": "Pinecone Vector Database", "category": "向量数据库", "description": "Pinecone在2026年保持托管型向量数据库领导地位，是AI Agent生产系统的首选方案，支持大规模向量检索和混合查询", "key_features": [ "全托管服务", "大规模向量检索", "混合查询", "Serverless架构" ], "status": "production", "year": 2026, "maintainer": "Pinecone", "language": "Rust/Go", "sources": [ { "source_type": "web", "source_credibility": "B", "article_url": "https://karthikeyanrathinam.medium.com/top-10-vector-databases-in-2026-ultimate-comparison-benchmarks-use-cases-6b0e878256b5", "collected_at": "2026-06-16T22:31:38.000Z" } ] }, { "id": "VDB-qdrant-2026", "name": "Qdrant 2026 Major Update", "full_name": "Qdrant Vector Search Engine", "category": "向量数据库", "description": "Qdrant在2026年发布重大更新，作为Rust编写的开源向量数据库速度领导者，在AI Agent工作负载中表现突出", "key_features": [ "Rust实现", "开源", "速度领先", "2026重大更新", "高效过滤" ], "status": "production", "year": 2026, "maintainer": "Qdrant", "language": "Rust", "sources": [ { "source_type": "web", "source_credibility": "B", "article_url": "https://karthikeyanrathinam.medium.com/top-10-vector-databases-in-2026-ultimate-comparison-benchmarks-use-cases-6b0e878256b5", "collected_at": "2026-06-16T22:31:38.000Z" } ] }, { "name": "Chroma", "type": "Vector Database", "description": "A modern vector database designed for AI and machine learning applications with ease of use and integration.", "features": [ "Developer-friendly API", "Seamless ML integration", "Lightweight deployment" ], "status": "Active", "sources": [ "https://www.datacamp.com/blog/the-top-5-vector-databases" ], "id": "VDB-036" }, { "name": "Oracle AI Vector Search", "type": "Vector Database Feature", "description": "Oracle's AI-powered vector search capability within Oracle AI Database 26ai, enabling fast similarity searches using AI-generated embeddings.", "features": [ "AI-generated vector embeddings", "Lightning-fast similarity search", "Integration with Oracle Database ecosystem" ], "status": "Active", "sources": [ "https://www.oracle.com/database/ai-vector-search" ], "id": "VDB-037" } ], "benchmarks": [ { "id": "BENCH-001", "name": "SWE-bench", "focus": "代码修复", "description": "真实GitHub issue自动修复", "top_score": "Claude 3.5 Sonnet ~49%", "adoption": "广泛" }, { "id": "BENCH-002", "name": "WebArena", "focus": "网页操作", "description": "真实网页环境任务完成", "top_score": "~35%", "adoption": "学术" }, { "id": "BENCH-003", "name": "AgentBench", "focus": "多任务", "description": "多场景Agent能力评估", "top_score": "GPT-4 ~36.9", "adoption": "学术" }, { "id": "BENCH-004", "name": "BFCL v3", "focus": "函数调用", "description": "多语言函数调用评估", "top_score": "GPT-4o ~90%", "adoption": "广泛" }, { "id": "BENCH-005", "name": "MCP兼容性测试", "focus": "MCP", "description": "MCP Server兼容性验证", "top_score": "N/A", "adoption": "新兴" }, { "id": "BENCH-006", "name": "LoCoMo", "type": "memory_benchmark", "description": "LoCoMo，2026年Agent记忆评估标准基准，用于比较不同记忆架构的长期对话能力", "category": "memory", "release_date": "2026", "source": "web_search_2026" }, { "id": "BENCH-007", "name": "LongMemEval", "type": "memory_benchmark", "description": "LongMemEval，2026年Agent长期记忆评估基准，测试Agent跨会话记忆召回和推理能力", "category": "memory", "release_date": "2026", "source": "web_search_2026" }, { "id": "BENCH-008", "name": "BEAM", "type": "memory_benchmark", "description": "BEAM，2026年Agent记忆评估基准，标准化Agent记忆系统性能比较框架", "category": "memory", "release_date": "2026", "source": "web_search_2026" }, { "name": "GAIA", "type": "Benchmark", "description": "A general-assistant benchmark measuring AI agents' ability to perform tasks requiring web search, file manipulation, and multi-step reasoning.", "metrics": [ "Task success rate", "Cost per evaluation" ], "status": "Active", "sources": [ "https://www.spheron.network/blog/ai-agent-benchmarking-gpu-cloud-swebench-gaia", "https://www.reddit.com/r/LocalLLaMA/comments/1hqt79i/top_agent_only_27_away_from_degreeholding_humans", "https://decodethefuture.org/en/ai-agent-benchmarks-2026" ], "id": "BENCH-009" }, { "name": "Terminal-Bench", "type": "Benchmark", "description": "A benchmark evaluating AI coding agents' capabilities in terminal-based interactions and command-line operations beyond traditional code repositories.", "metrics": [ "Command execution accuracy", "Terminal interaction success" ], "status": "Active", "sources": [ "https://www.birjob.com/blog/agent-benchmarks-2026" ], "id": "BENCH-010" } ], "sdks": [ { "id": "SDK-001", "name": "Claude Agent SDK", "vendor": "Anthropic", "language": "Python", "type": "Agent SDK", "multi_agent": true, "tool_use": true, "memory": true, "streaming": true, "maturity": "生产可用", "features": [ "原生MCP支持", "多步骤agent", "guardrails" ], "pros": [ "Anthropic官方", "MCP原生集成", "安全护栏" ], "cons": [ "仅Claude模型", "较新生态小" ] }, { "id": "SDK-002", "name": "OpenAI Agents SDK", "vendor": "OpenAI", "language": "Python", "type": "Agent SDK", "multi_agent": true, "tool_use": true, "memory": false, "streaming": true, "maturity": "生产可用", "features": [ "Agent交接", "guardrails", "tracing" ], "pros": [ "OpenAI官方", "Agent交接优雅", "内置tracing" ], "cons": [ "仅OpenAI模型", "无内置记忆" ] }, { "id": "SDK-003", "name": "Google ADK", "vendor": "Google", "language": "Python", "type": "Agent SDK", "multi_agent": true, "tool_use": true, "memory": true, "streaming": true, "maturity": "早期", "features": [ "A2A协议原生", "多Agent编排", "内置评估" ], "pros": [ "A2A原生", "Google生态", "多模型支持" ], "cons": [ "非常新", "文档少", "生态小" ] }, { "id": "SDK-004", "name": "LangGraph", "vendor": "LangChain", "language": "Python/JS", "type": "Agent框架", "multi_agent": true, "tool_use": true, "memory": true, "streaming": true, "maturity": "生产可用", "features": [ "状态图", "持久化", "人机协作", "时间旅行调试" ], "pros": [ "最成熟", "社区大", "可视化调试" ], "cons": [ "学习曲线陡", "复杂度高", "性能开销大" ] }, { "id": "SDK-005", "name": "CrewAI", "vendor": "CrewAI", "language": "Python", "type": "Agent框架", "multi_agent": true, "tool_use": true, "memory": true, "streaming": true, "maturity": "生产可用", "features": [ "角色定义", "任务委派", "流程编排", "记忆系统" ], "pros": [ "概念直观", "上手快", "角色扮演模式" ], "cons": [ "灵活性有限", "调试困难", "企业版收费" ] }, { "id": "SDK-006", "name": "AutoGen/AG2", "vendor": "Microsoft", "language": "Python", "type": "Agent框架", "multi_agent": true, "tool_use": true, "memory": true, "streaming": true, "maturity": "生产可用", "features": [ "对话式多Agent", "代码执行", "人机协作" ], "pros": [ "微软背书", "研究导向", "灵活对话模式" ], "cons": [ "API不稳定", "文档混乱", "版本迁移痛苦" ] }, { "id": "SDK-007", "name": "Smolagents", "vendor": "HuggingFace", "language": "Python", "type": "Agent框架", "multi_agent": false, "tool_use": true, "memory": false, "streaming": true, "maturity": "早期", "features": [ "轻量级", "代码Agent", "HF模型集成" ], "pros": [ "极简", "HF生态", "代码优先" ], "cons": [ "单Agent", "功能少", "生态小" ] }, { "id": "SDK-008", "name": "Pydantic AI", "vendor": "Pydantic", "language": "Python", "type": "Agent框架", "multi_agent": false, "tool_use": true, "memory": false, "streaming": true, "maturity": "早期", "features": [ "类型安全", "依赖注入", "多模型支持" ], "pros": [ "类型安全", "Pydantic生态", "模型无关" ], "cons": [ "单Agent", "无记忆", "新项目" ] }, { "id": "SDK-009", "name": "Mastra", "vendor": "Mastra", "language": "TypeScript", "type": "Agent框架", "multi_agent": true, "tool_use": true, "memory": true, "streaming": true, "maturity": "早期", "features": [ "TypeScript原生", "工作流", "RAG集成" ], "pros": [ "TS开发者友好", "全栈方案" ], "cons": [ "TS only", "新项目", "社区小" ] }, { "id": "SDK-010", "name": "Dify", "vendor": "LangGenius", "language": "Python/TS", "type": "Agent平台", "multi_agent": true, "tool_use": true, "memory": true, "streaming": true, "maturity": "生产可用", "features": [ "可视化编排", "插件市场", "RAG", "工作流" ], "pros": [ "低代码", "可视化", "插件生态", "中文友好" ], "cons": [ "定制性差", "性能开销", "复杂场景受限" ] }, { "id": "SDK-011", "name": "Semantic Kernel", "vendor": "Microsoft", "language": "C#/Python/Java", "type": "Agent SDK", "multi_agent": true, "tool_use": true, "memory": true, "streaming": true, "maturity": "生产可用", "features": [ "企业级", "多语言", "插件系统", "流程编排" ], "pros": [ "企业级支持", "多语言", "Azure集成", "成熟稳定" ], "cons": [ "C#优先", "学习曲线陡", "Python支持弱于C#" ] }, { "id": "SDK-012", "name": "Phidata", "vendor": "Phidata", "language": "Python", "type": "Agent框架", "multi_agent": true, "tool_use": true, "memory": true, "streaming": true, "maturity": "生产可用", "features": [ "Agent类", "工具集成", "知识库", "多模型" ], "pros": [ "简单易用", "内置知识库", "快速原型" ], "cons": [ "社区小", "定制性有限", "文档不完善" ] }, { "id": "SDK-013", "name": "LlamaIndex", "vendor": "LlamaIndex", "language": "Python/TS", "type": "Agent框架", "multi_agent": true, "tool_use": true, "memory": true, "streaming": true, "maturity": "生产可用", "features": [ "RAG专精", "数据连接器", "Agent工作流", "查询引擎" ], "pros": [ "RAG最强", "数据连接器丰富", "文档好" ], "cons": [ "RAG偏重", "Agent功能较新", "性能开销" ] }, { "id": "SDK-014", "name": "Haystack", "vendor": "deepset", "language": "Python", "type": "Agent框架", "multi_agent": true, "tool_use": true, "memory": true, "streaming": true, "maturity": "生产可用", "features": [ "Pipeline架构", "RAG", "搜索", "Agent" ], "pros": [ "Pipeline灵活", "生产就绪", "搜索专精" ], "cons": [ "Agent功能较新", "社区小于LangChain", "学习曲线" ] }, { "id": "SDK-015", "name": "Agency Swarm", "vendor": "VRSEN", "language": "Python", "type": "Agent框架", "multi_agent": true, "tool_use": true, "memory": true, "streaming": true, "maturity": "早期", "features": [ "Agent层级", "角色化", "自动工具创建" ], "pros": [ "层级结构清晰", "自动工具生成", "OpenAI原生" ], "cons": [ "仅OpenAI", "社区小", "文档少" ] }, { "id": "SDK-016", "name": "TaskWeaver", "vendor": "Microsoft", "language": "Python", "type": "Agent框架", "multi_agent": false, "tool_use": true, "memory": true, "streaming": true, "maturity": "早期", "features": [ "代码解释器", "数据分析", "插件系统" ], "pros": [ "数据分析专精", "代码执行安全", "微软支持" ], "cons": [ "单Agent", "数据分析偏重", "社区小" ] }, { "id": "SDK-017", "name": "Camel-AI", "vendor": "Camel-AI", "language": "Python", "type": "Agent框架", "multi_agent": true, "tool_use": true, "memory": true, "streaming": true, "maturity": "早期", "features": [ "角色扮演", "多Agent协作", "社会模拟" ], "pros": [ "角色扮演创新", "研究导向", "多Agent交互" ], "cons": [ "研究阶段", "生产就绪度低", "文档少" ] }, { "id": "SDK-018", "name": "Vercel AI SDK", "vendor": "Vercel", "language": "TypeScript", "type": "Agent SDK", "multi_agent": true, "tool_use": true, "memory": true, "streaming": true, "maturity": "生产可用", "features": [ "全栈AI", "流式UI", "多模型", "工具调用" ], "pros": [ "Next.js集成", "流式渲染", "多模型支持", "生产就绪" ], "cons": [ "TS only", "前端偏重", "复杂Agent逻辑受限" ] }, { "id": "SDK-019", "name": "Rig", "vendor": "0xPlaygrounds", "language": "Rust", "type": "Agent框架", "multi_agent": true, "tool_use": true, "memory": true, "streaming": true, "maturity": "早期", "features": [ "Rust原生", "高性能", "Agent编排" ], "pros": [ "极致性能", "类型安全", "Rust生态" ], "cons": [ "Rust only", "极新", "社区极小" ] }, { "id": "SDK-020", "name": "Portkey AI Gateway", "vendor": "Portkey", "language": "Python/TS", "type": "Agent网关", "multi_agent": true, "tool_use": true, "memory": false, "streaming": true, "maturity": "生产可用", "features": [ "多模型路由", "可观测性", "Fallback", "缓存" ], "pros": [ "模型无关", "生产级网关", "监控完善" ], "cons": [ "网关非框架", "无Agent逻辑", "付费功能" ] }, { "id": "SDK-021", "name": "Coze", "vendor": "字节跳动", "language": "Python/TS", "type": "Agent平台", "multi_agent": true, "tool_use": true, "memory": true, "streaming": true, "maturity": "生产可用", "features": [ "可视化编排", "插件市场", "多模型", "工作流" ], "pros": [ "中文友好", "低代码", "字节生态", "免费额度" ], "cons": [ "定制性差", "平台锁定", "复杂场景受限" ] }, { "id": "SDK-022", "name": "Flowise", "vendor": "FlowiseAI", "language": "TypeScript", "type": "Agent平台", "multi_agent": true, "tool_use": true, "memory": true, "streaming": true, "maturity": "生产可用", "features": [ "拖拽式编排", "LangChain集成", "API导出" ], "pros": [ "零代码", "可视化", "快速原型" ], "cons": [ "定制性差", "性能开销", "复杂逻辑受限" ] }, { "id": "SDK-023", "name": "n8n AI Agent", "vendor": "n8n", "language": "TypeScript", "type": "Agent平台", "multi_agent": true, "tool_use": true, "memory": true, "streaming": true, "maturity": "生产可用", "features": [ "工作流自动化", "AI节点", "400+集成", "自托管" ], "pros": [ "集成最多", "自托管", "可视化", "开源" ], "cons": [ "Agent功能新", "AI深度有限", "复杂推理受限" ] }, { "id": "SDK-024", "name": "AgentScope", "vendor": "Alibaba", "language": "Python", "type": "Agent框架", "multi_agent": true, "tool_use": true, "memory": true, "streaming": true, "maturity": "早期", "features": [ "分布式Agent", "消息机制", "服务集成" ], "pros": [ "分布式设计", "阿里背书", "中文文档" ], "cons": [ "新项目", "社区小", "文档不完善" ] }, { "id": "SDK-025", "name": "Bee Agent Framework", "vendor": "IBM", "language": "TypeScript", "type": "Agent框架", "multi_agent": true, "tool_use": true, "memory": true, "streaming": true, "maturity": "早期", "features": [ "企业级", "可观测性", "多模型", "工具集成" ], "pros": [ "IBM背书", "企业级设计", "TypeScript" ], "cons": [ "极新", "社区小", "文档少" ] }, { "id": "SDK-026", "name": "AutoGen Studio", "vendor": "Microsoft", "language": "Python", "type": "Agent平台", "multi_agent": true, "tool_use": true, "memory": true, "streaming": true, "maturity": "早期", "features": [ "可视化AutoGen", "拖拽编排", "Web UI" ], "pros": [ "AutoGen可视化", "微软支持", "快速原型" ], "cons": [ "功能有限", "不稳定", "生产就绪度低" ] }, { "id": "SDK-027", "name": "Instructor", "vendor": "Jason Liu", "language": "Python/TS", "type": "Agent工具库", "multi_agent": false, "tool_use": true, "memory": false, "streaming": true, "maturity": "生产可用", "features": [ "结构化输出", "Pydantic验证", "重试机制" ], "pros": [ "结构化输出最强", "简单可靠", "多模型支持" ], "cons": [ "非Agent框架", "无编排", "工具库定位" ] }, { "id": "SDK-028", "name": "Mirage", "vendor": "Mirage", "language": "Python", "type": "Agent框架", "multi_agent": true, "tool_use": true, "memory": true, "streaming": true, "maturity": "早期", "features": [ "多Agent沙箱", "安全执行", "代码Agent" ], "pros": [ "安全沙箱", "代码执行安全", "多Agent" ], "cons": [ "极新", "社区小", "文档少" ] }, { "id": "SDK-029", "name": "DSPy", "vendor": "Stanford", "language": "Python", "type": "Agent框架", "multi_agent": false, "tool_use": true, "memory": false, "streaming": true, "maturity": "早期", "features": [ "声明式编程", "自动优化", "模块化" ], "pros": [ "学术前沿", "自动Prompt优化", "模块化设计" ], "cons": [ "学习曲线陡", "生产就绪度低", "社区小" ] }, { "id": "SDK-030", "name": "Fixpoint/Wordware", "vendor": "Fixpoint", "language": "Python/TS", "type": "Agent平台", "multi_agent": true, "tool_use": true, "memory": true, "streaming": true, "maturity": "早期", "features": [ "可视化编排", "自然语言编程", "API导出" ], "pros": [ "自然语言编程", "可视化", "快速迭代" ], "cons": [ "新项目", "社区小", "定制性有限" ] }, { "id": "SDK-031", "name": "Letta (MemGPT)", "vendor": "Letta", "language": "Python", "type": "Agent框架", "multi_agent": true, "tool_use": true, "memory": true, "streaming": true, "maturity": "生产可用", "features": [ "虚拟上下文管理", "无限记忆", "自动分页", "状态管理" ], "pros": [ "记忆管理最强", "无限上下文", "研究背书" ], "cons": [ "复杂度高", "部署要求高", "社区中等" ] }, { "id": "SDK-032", "name": "OpenRouter SDK", "vendor": "OpenRouter", "language": "Python/TS", "type": "Agent网关", "multi_agent": false, "tool_use": true, "memory": false, "streaming": true, "maturity": "生产可用", "features": [ "100+模型路由", "统一API", "价格比较" ], "pros": [ "模型最多", "统一接口", "价格透明" ], "cons": [ "网关非框架", "延迟较高", "无Agent逻辑" ] }, { "id": "SDK-openai-agents", "name": "OpenAI Agents SDK", "type": "agent_framework", "company": "OpenAI", "language": "Python", "license": "MIT", "key_features": [ "Released March 2025, replaced experimental Swarm framework", "Production-grade toolkit with handoff pattern", "Tight integration with OpenAI Responses API and function calling", "Native MCP support", "Guardrails and tracing built-in", "Best for OpenAI-native workloads" ], "status": "production", "year": 2025, "sources": [ { "source_type": "web", "source_credibility": "A", "collected_at": "2026-05-29T14:25:59.103Z" } ] }, { "id": "SDK-ms-agent-framework", "name": "Microsoft Agent Framework 1.0", "type": "agent_framework", "company": "Microsoft", "language": "Python/.NET", "license": "MIT", "key_features": [ "GA release April 3, 2026", "Unified AutoGen + Semantic Kernel into single production SDK", "Stable APIs across .NET and Python", "GraphFlow for explicit workflow control", "AgentSession for persistent conversation state", "MCP server support for standardized tool integration", "Hosted tools for Azure-native execution" ], "status": "production", "year": 2026, "sources": [ { "source_type": "web", "source_credibility": "A", "collected_at": "2026-05-29T14:25:59.103Z" } ] }, { "id": "SDK-mastra", "name": "Mastra", "type": "agent_framework", "company": "Mastra", "language": "TypeScript", "license": "MIT", "key_features": [ "TypeScript-first AI agent framework", "Built-in RAG, evals, and observability", "Production-ready with deployment tooling" ], "status": "production", "year": 2025, "sources": [ { "source_type": "web", "source_credibility": "B", "collected_at": "2026-05-29T14:25:59.103Z" } ] }, { "id": "SDK-pydantic-ai", "name": "Pydantic AI", "type": "agent_framework", "company": "Pydantic Team", "language": "Python", "license": "MIT", "key_features": [ "Type-safe agent framework built on Pydantic", "Model-agnostic with structured output", "Zero overhead (0-10ms per call)", "Best for teams wanting minimal framework lock-in" ], "status": "production", "year": 2025, "sources": [ { "source_type": "web", "source_credibility": "B", "collected_at": "2026-05-29T14:25:59.103Z" } ] }, { "id": "SDK-smolagents", "name": "Smolagents", "type": "agent_framework", "company": "Hugging Face", "language": "Python", "license": "Apache-2.0", "key_features": [ "Lightweight agent framework from Hugging Face", "Code-based agent actions (not just tool calls)", "Minimal abstraction, easy to understand" ], "status": "production", "year": 2025, "sources": [ { "source_type": "web", "source_credibility": "B", "collected_at": "2026-05-29T14:25:59.103Z" } ] }, { "id": "SDK-hermes", "name": "Hermes Agent", "type": "agent_framework", "company": "Hermes AI", "language": "Python/TypeScript", "license": "Apache-2.0", "key_features": [ "从经验中创建技能", "使用中改进技能", "构建用户模型", "长期自主运行", "安全架构" ], "status": "production", "year": 2026, "sources": [ { "source_type": "web", "source_credibility": "B", "collected_at": "2026-05-30T13:38:30.808Z" } ] }, { "id": "SDK-openclaw", "name": "OpenClaw", "type": "agent_framework", "company": "开源社区", "language": "Python", "license": "MIT", "key_features": [ "最广泛的集成支持", "社区构建技能", "多渠道支持", "开源Agent平台" ], "status": "production", "year": 2025, "sources": [ { "source_type": "web", "source_credibility": "B", "collected_at": "2026-05-30T13:38:30.808Z" } ] }, { "id": "SDK-deerflow", "name": "DeerFlow", "type": "agent_framework", "company": "ByteDance", "language": "Python", "license": "Apache-2.0", "key_features": [ "多Agent协作", "研究导向", "深度搜索集成", "报告生成" ], "status": "production", "year": 2025, "sources": [ { "source_type": "web", "source_credibility": "B", "collected_at": "2026-05-30T13:38:30.808Z" } ] }, { "id": "SDK-agno", "name": "AGNO (Phidata v2)", "type": "agent_framework", "company": "AGNO", "language": "Python", "license": "MIT", "key_features": [ "轻量级Agent框架", "多模型支持", "工具集成", "Agent团队协作" ], "status": "production", "year": 2025, "sources": [ { "source_type": "web", "source_credibility": "B", "collected_at": "2026-05-30T13:38:30.808Z" } ] }, { "id": "SDK-mem0", "name": "Mem0 SDK", "type": "memory_sdk", "company": "Mem0", "language": "Python/TypeScript", "license": "MIT", "key_features": [ "AI Agent记忆层", "图记忆", "自动记忆提取", "跨会话持久化", "55K+ GitHub Stars" ], "status": "production", "year": 2025, "sources": [ { "source_type": "web", "source_credibility": "A", "collected_at": "2026-05-30T13:38:30.808Z" } ] }, { "id": "SDK-zep", "name": "Zep", "type": "memory_sdk", "company": "Zep", "language": "Python/TypeScript", "license": "MIT", "key_features": [ "时序关系记忆", "知识图谱", "对话历史管理", "结构化记忆提取" ], "status": "production", "year": 2025, "sources": [ { "source_type": "web", "source_credibility": "B", "collected_at": "2026-05-30T13:38:30.808Z" } ] }, { "id": "SDK-langmem", "name": "LangMem", "type": "memory_sdk", "company": "LangChain", "language": "Python", "license": "MIT", "key_features": [ "LangGraph原生记忆", "工作流编排集成", "记忆工具链", "语义搜索" ], "status": "production", "year": 2025, "sources": [ { "source_type": "web", "source_credibility": "B", "collected_at": "2026-05-30T13:38:30.808Z" } ] }, { "id": "SDK-hindsight", "name": "Hindsight", "type": "memory_sdk", "company": "Hindsight AI", "language": "Python", "license": "MIT", "key_features": [ "记忆准确率91.4%", "自动记忆管理", "上下文感知检索", "轻量级部署" ], "status": "production", "year": 2026, "sources": [ { "source_type": "web", "source_credibility": "B", "collected_at": "2026-05-30T13:38:30.808Z" } ] }, { "id": "SDK-cognee", "name": "Cognee", "type": "memory_sdk", "company": "Cognee", "language": "Python", "license": "Apache-2.0", "key_features": [ "认知记忆架构", "知识图谱构建", "多模态记忆", "可扩展存储" ], "status": "production", "year": 2025, "sources": [ { "source_type": "web", "source_credibility": "B", "collected_at": "2026-05-30T13:38:30.808Z" } ] }, { "id": "SDK-memvid", "name": "Memvid", "type": "memory_sdk", "company": "开源社区", "language": "Python", "license": "MIT", "key_features": [ "视频记忆编码", "多模态记忆检索", "视觉上下文保存" ], "status": "experimental", "year": 2026, "sources": [ { "source_type": "web", "source_credibility": "C", "collected_at": "2026-05-30T13:38:30.808Z" } ] }, { "id": "SDK-memoclaw", "name": "MemoClaw", "type": "memory_sdk", "company": "MemoClaw", "language": "Python/TypeScript", "license": "MIT", "key_features": [ "Agent记忆管理", "跨平台记忆同步", "结构化记忆存储" ], "status": "production", "year": 2026, "sources": [ { "source_type": "web", "source_credibility": "C", "collected_at": "2026-05-30T13:38:30.808Z" } ] }, { "id": "SDK-evermind", "name": "EverMind", "type": "memory_sdk", "company": "EverMind AI", "language": "Python", "license": "MIT", "key_features": [ "Zep替代方案", "长期记忆管理", "对话理解", "个性化记忆" ], "status": "production", "year": 2026, "sources": [ { "source_type": "web", "source_credibility": "C", "collected_at": "2026-05-30T13:38:30.808Z" } ] }, { "id": "SDK-fastmcp", "name": "FastMCP SDK", "type": "mcp_sdk", "company": "FastMCP", "language": "Python", "license": "MIT", "key_features": [ "快速MCP Server开发", "类型安全", "自动文档生成", "热重载" ], "status": "production", "year": 2025, "sources": [ { "source_type": "web", "source_credibility": "B", "collected_at": "2026-05-30T13:38:30.808Z" } ] }, { "id": "SDK-crewai-v2", "name": "CrewAI v2", "type": "agent_framework", "company": "CrewAI", "language": "Python", "license": "MIT", "key_features": [ "多Agent协作", "角色定义", "任务委派", "流程编排", "企业版支持" ], "status": "production", "year": 2026, "sources": [ { "source_type": "web", "source_credibility": "B", "collected_at": "2026-05-30T13:38:30.808Z" } ] }, { "id": "SDK-autogen-v2", "name": "AutoGen v0.5 (AG2)", "type": "agent_framework", "company": "Microsoft", "language": "Python", "license": "MIT", "key_features": [ "多Agent对话", "代码执行", "人机协作", "可组合Agent" ], "status": "production", "year": 2025, "sources": [ { "source_type": "web", "source_credibility": "A", "collected_at": "2026-05-30T13:38:30.808Z" } ] }, { "id": "SDK-semantic-kernel-v2", "name": "Semantic Kernel v2", "type": "agent_sdk", "company": "Microsoft", "language": "C#/Python/Java", "license": "MIT", "key_features": [ "企业级Agent SDK", "多语言支持", "插件架构", "AI编排" ], "status": "production", "year": 2025, "sources": [ { "source_type": "web", "source_credibility": "A", "collected_at": "2026-05-30T13:38:30.808Z" } ] }, { "id": "SDK-bee-v2", "name": "Bee Agent Framework v2", "type": "agent_framework", "company": "IBM", "language": "TypeScript", "license": "Apache-2.0", "key_features": [ "IBM官方Agent框架", "TypeScript原生", "企业集成", "ACP协议支持" ], "status": "production", "year": 2026, "sources": [ { "source_type": "web", "source_credibility": "A", "collected_at": "2026-05-30T13:38:30.808Z" } ] }, { "id": "SDK-inngest", "name": "Inngest", "type": "agent_orchestration", "company": "Inngest", "language": "TypeScript/Python", "license": "MIT", "key_features": [ "Durable execution for AI agents", "Event-driven workflow orchestration", "Automatic retries and timeouts", "Step functions for long-running agents", "Production-grade reliability" ], "status": "production", "year": 2026, "sources": [ { "source_type": "web", "source_credibility": "B", "collected_at": "2026-05-31T13:41:37.338Z" } ] }, { "id": "SDK-trigger-dev", "name": "Trigger.dev", "type": "agent_orchestration", "company": "Trigger.dev", "language": "TypeScript", "license": "Apache-2.0", "key_features": [ "TypeScript-native AI workflow platform", "Durable tasks with automatic retries", "Background AI jobs", "Real-time streaming", "Integration with OpenAI/Anthropic" ], "status": "production", "year": 2026, "sources": [ { "source_type": "web", "source_credibility": "B", "collected_at": "2026-05-31T13:41:37.338Z" } ] }, { "id": "SDK-temporal-agents", "name": "Temporal Agent Workflow", "type": "agent_orchestration", "company": "Temporal", "language": "TypeScript/Go/Python", "license": "MIT", "key_features": [ "Stateful workflow engine for AI agents", "Durable execution with exact-once semantics", "Long-running agent orchestration", "Built-in observability and debugging", "Scales to millions of concurrent workflows" ], "status": "production", "year": 2026, "sources": [ { "source_type": "web", "source_credibility": "A", "collected_at": "2026-05-31T13:41:37.338Z" } ] }, { "id": "SDK-google-a2a", "name": "Google A2A SDK", "type": "agent_protocol_sdk", "company": "Google", "language": "Python/TypeScript", "license": "Apache-2.0", "key_features": [ "Official SDK for Agent-to-Agent protocol", "Agent Card discovery and registration", "Task-based communication", "150+ organizations adopted", "Native integration with Google ADK" ], "status": "production", "year": 2026, "sources": [ { "source_type": "web", "source_credibility": "A", "collected_at": "2026-05-31T13:41:37.338Z" } ] }, { "id": "SDK-openai-responses", "name": "OpenAI Responses API SDK", "type": "agent_sdk", "company": "OpenAI", "language": "Python/TypeScript", "license": "MIT", "key_features": [ "New Responses API for agent interactions", "Built-in tool use and function calling", "Streaming support", "MCP integration", "Replaces Chat Completions for agents" ], "status": "production", "year": 2026, "sources": [ { "source_type": "web", "source_credibility": "A", "collected_at": "2026-05-31T13:41:37.338Z" } ] }, { "id": "SDK-langgraph-platform", "name": "LangGraph Platform", "type": "agent_platform", "company": "LangChain", "language": "Python/TypeScript", "license": "MIT", "key_features": [ "Managed LangGraph deployment platform", "Built-in persistence and memory", "Cron triggers and webhooks", "Real-time streaming", "Human-in-the-loop support" ], "status": "production", "year": 2026, "sources": [ { "source_type": "web", "source_credibility": "A", "collected_at": "2026-05-31T13:41:37.338Z" } ] }, { "id": "SDK-scrapling", "name": "Scrapling", "category": "sdks", "type": "Web Scraping Framework", "language": "Python", "version": "0.4.8", "github": "https://github.com/D4Vinci/Scrapling", "pypi": "scrapling", "description": "Adaptive web scraping framework with anti-bot bypass, smart element tracking, and AI agent integration", "key_features": [ "Adaptive element selection (no fragile selectors)", "Native Cloudflare/anti-bot bypass", "Three-tier fetcher system", "Built-in Spider framework", "MCP Server for AI agents", "Auto-resuming crawls" ], "installation": "pip install scrapling", "license": "BSD-3", "sources": [ "web_search", "github" ], "confidence": "high", "first_seen": "2026-05-31T22:02:40.222Z", "last_updated": "2026-05-31T22:02:40.222Z" }, { "id": "SDK-hermes-v2", "name": "Hermes Agent v2", "type": "agent_framework", "company": "Hermes AI", "language": "Python/TypeScript", "license": "Apache-2.0", "key_features": [ "从经验中创建技能", "使用中改进技能", "构建用户模型", "长期自主运行", "Forbes报道超越OpenClaw", "企业级安全架构" ], "status": "production", "year": 2026, "sources": [ { "source_type": "web", "source_credibility": "A", "collected_at": "2026-06-01T13:48:20.198Z" } ] }, { "id": "SDK-pickaxe", "name": "Pickaxe AI Agent Framework", "type": "agent_framework", "company": "Pickaxe", "language": "Python/TypeScript", "license": "MIT", "key_features": [ "2026 Top 15 AI Agent框架", "多模型支持", "快速部署", "可视化编排" ], "status": "production", "year": 2026, "sources": [ { "source_type": "web", "source_credibility": "B", "collected_at": "2026-06-01T13:48:20.198Z" } ] }, { "id": "SDK-flowtivity", "name": "Flowtivity", "type": "agent_orchestration", "company": "Flowtivity", "language": "Python", "license": "MIT", "key_features": [ "Agent工作流编排", "Hermes/OpenClaw对比", "预算优化模型", "企业集成" ], "status": "production", "year": 2026, "sources": [ { "source_type": "web", "source_credibility": "B", "collected_at": "2026-06-01T13:48:20.198Z" } ] }, { "id": "SDK-akka-agents", "name": "Akka Agent Runtime", "type": "agent_orchestration", "company": "Akka", "language": "TypeScript/Java/Scala", "license": "Apache-2.0", "key_features": [ "分布式Agent运行时", "事件驱动架构", "Inngest vs Temporal对比", "云原生部署" ], "status": "production", "year": 2026, "sources": [ { "source_type": "web", "source_credibility": "B", "collected_at": "2026-06-01T13:48:20.198Z" } ] }, { "id": "SDK-vellum", "name": "Vellum AI", "type": "agent_platform", "company": "Vellum", "language": "Python/TypeScript", "license": "Commercial", "key_features": [ "Top 11 AI Agent框架2026", "LLM编排", "Prompt管理", "工作流测试" ], "status": "production", "year": 2026, "sources": [ { "source_type": "web", "source_credibility": "B", "collected_at": "2026-06-01T13:48:20.198Z" } ] }, { "id": "SDK-gurusup", "name": "GuruSup Multi-Agent", "type": "agent_framework", "company": "GuruSup", "language": "Python", "license": "MIT", "key_features": [ "2026最佳多Agent框架", "OpenAI Agents SDK集成", "CrewAI对比", "生产部署指南" ], "status": "production", "year": 2026, "sources": [ { "source_type": "web", "source_credibility": "B", "collected_at": "2026-06-01T13:48:20.198Z" } ] }, { "id": "SDK-vercel-ai-v2", "name": "Vercel AI SDK v2", "type": "agent_framework", "description": "Vercel AI SDK v2，2026年更新版本，支持多模型Agent编排、流式响应和边缘部署", "vendor": "Vercel", "language": "TypeScript", "github_stars": "15000+", "release_date": "2026", "source": "web_search_2026" }, { "id": "SDK-crewai-enterprise", "name": "CrewAI Enterprise", "type": "agent_platform", "description": "CrewAI企业版，支持多Agent协作、企业级安全和监控，2026年推出", "vendor": "CrewAI", "language": "Python", "release_date": "2026", "source": "web_search_2026" }, { "id": "SDK-langgraph-studio", "name": "LangGraph Studio", "type": "agent_platform", "description": "LangGraph Studio可视化Agent开发平台，支持拖拽式Agent工作流设计和调试", "vendor": "LangChain", "language": "Python/TypeScript", "release_date": "2026", "source": "web_search_2026" }, { "id": "SDK-google-a2a-v2", "name": "Google A2A SDK v2", "type": "agent_protocol_sdk", "description": "Google A2A SDK v2，支持A2A v1协议，15000+ GitHub stars，多Agent间安全通信和协作", "vendor": "Google", "language": "Python/TypeScript", "github_stars": "15000+", "release_date": "2026", "source": "web_search_2026" }, { "id": "SDK-openai-agents-v2", "name": "OpenAI Agents SDK v2", "type": "agent_framework", "description": "OpenAI Agents SDK v2，19000+ GitHub stars，1030万月下载量，生产级多Agent框架", "vendor": "OpenAI", "language": "Python", "github_stars": "19000+", "monthly_downloads": "10.3M", "release_date": "2026", "source": "web_search_2026" }, { "id": "SDK-claude-agent-v2", "name": "Claude Agent SDK v2", "type": "agent_framework", "description": "Anthropic Claude Agent SDK v2，2026年增长最快的Agent框架之一，原生支持Claude模型", "vendor": "Anthropic", "language": "Python", "release_date": "2026", "source": "web_search_2026" }, { "id": "SDK-ms-maf-v2", "name": "Microsoft Agent Framework v2", "type": "agent_framework", "description": "Microsoft Agent Framework v2，支持.NET和Python多语言Agent开发，生产级多Agent工作流", "vendor": "Microsoft", "language": ".NET/Python", "release_date": "2026", "source": "web_search_2026" }, { "id": "sdk-mastra-v2", "name": "Mastra v2", "type": "agent_framework", "description": "Mastra v2 TypeScript原生Agent框架，专为构建AI Agent和编排工作流设计，高度专业化", "company": "Mastra", "language": "TypeScript", "license": "Apache-2.0", "key_features": [ "TypeScript原生", "Agent编排", "工作流引擎", "RAG集成", "MCP支持" ], "status": "production", "year": 2026, "sources": [ { "source_type": "web", "source_credibility": "B", "collected_at": "2026-06-03T13:52:48.985531Z" } ] }, { "id": "sdk-deerflow-v2", "name": "DeerFlow v2", "type": "agent_framework", "description": "DeerFlow v2多Agent编排框架，支持复杂工作流和Agent协作", "company": "ByteDance", "language": "Python", "license": "Apache-2.0", "key_features": [ "多Agent编排", "工作流引擎", "Agent协作", "可视化调试" ], "status": "production", "year": 2026, "sources": [ { "source_type": "web", "source_credibility": "B", "collected_at": "2026-06-03T13:52:48.985531Z" } ] }, { "id": "sdk-agno-v2", "name": "AGNO v2 (Phidata v3)", "type": "agent_framework", "description": "AGNO v2（原Phidata），轻量级Agent框架，支持多模型和工具集成", "company": "AGNO", "language": "Python", "license": "MIT", "key_features": [ "轻量级", "多模型支持", "工具集成", "Agent工作流" ], "status": "production", "year": 2026, "sources": [ { "source_type": "web", "source_credibility": "B", "collected_at": "2026-06-03T13:52:48.985531Z" } ] }, { "id": "sdk-llamaindex-v2", "name": "LlamaIndex v2", "type": "rag_framework", "description": "LlamaIndex v2，RAG框架升级版，增强Agent能力和数据处理", "company": "LlamaIndex", "language": "Python/TypeScript", "license": "MIT", "key_features": [ "RAG增强", "Agent框架", "数据处理管道", "多模态支持" ], "status": "production", "year": 2026, "sources": [ { "source_type": "web", "source_credibility": "A", "collected_at": "2026-06-03T13:52:48.985531Z" } ] }, { "id": "sdk-flowise-v2", "name": "Flowise v2", "type": "visual_agent_builder", "description": "Flowise v2可视化Agent构建器，拖拽式Agent工作流设计", "company": "Flowise", "language": "TypeScript/Python", "license": "Apache-2.0", "key_features": [ "可视化构建", "拖拽式设计", "Agent工作流", "LLM集成" ], "status": "production", "year": 2026, "sources": [ { "source_type": "web", "source_credibility": "B", "collected_at": "2026-06-03T13:52:48.985531Z" } ] }, { "id": "sdk-n8n-v2", "name": "n8n AI Agent v2", "type": "workflow_automation", "description": "n8n AI Agent v2，工作流自动化平台升级版，增强AI Agent节点和工具集成", "company": "n8n", "language": "TypeScript", "license": "Sustainable Use", "key_features": [ "工作流自动化", "AI Agent节点", "400+集成", "自托管" ], "status": "production", "year": 2026, "sources": [ { "source_type": "web", "source_credibility": "B", "collected_at": "2026-06-03T13:52:48.985531Z" } ] }, { "id": "sdk-dify-v2", "name": "Dify v2", "type": "llm_app_platform", "description": "Dify v2 LLM应用开发平台，增强Agent能力和RAG管道", "company": "Dify", "language": "Python/TypeScript", "license": "Apache-2.0", "key_features": [ "LLM应用平台", "Agent工作流", "RAG管道", "可视化编排" ], "status": "production", "year": 2026, "sources": [ { "source_type": "web", "source_credibility": "B", "collected_at": "2026-06-03T13:52:48.985531Z" } ] }, { "id": "sdk-composio", "name": "Composio Agent SDK", "type": "tool_integration", "description": "Composio Agent SDK，250+工具集成中间件，为Agent提供工具调用能力", "company": "Composio", "language": "Python/TypeScript", "license": "MIT", "key_features": [ "250+工具集成", "OAuth管理", "Agent工具调用", "多框架兼容" ], "status": "production", "year": 2026, "sources": [ { "source_type": "web", "source_credibility": "B", "collected_at": "2026-06-03T13:52:48.985531Z" } ] }, { "id": "SDK-rasa-calm", "name": "Rasa CALM", "vendor": "Rasa", "language": "Python", "type": "Agent框架", "multi_agent": true, "tool_use": true, "memory": true, "streaming": true, "maturity": "生产可用", "features": [ "企业级对话AI", "CALM架构", "对话管理", "NLU引擎" ], "pros": [ "企业级", "对话AI专精", "成熟稳定" ], "cons": [ "对话AI偏重", "非通用Agent", "学习曲线" ], "sources": [ { "source_type": "web", "source_credibility": "B", "collected_at": "2026-06-04T14:03:15.064Z" } ] }, { "id": "SDK-morph-llm", "name": "Morph LLM Agent", "vendor": "Morph", "language": "Python/TypeScript", "type": "Agent框架", "multi_agent": true, "tool_use": true, "memory": false, "streaming": true, "maturity": "早期", "features": [ "8框架对比", "ACP协议支持", "轻量级", "快速推理" ], "pros": [ "轻量级", "ACP原生", "快速" ], "cons": [ "新项目", "社区小", "功能有限" ], "sources": [ { "source_type": "web", "source_credibility": "B", "collected_at": "2026-06-04T14:03:15.064Z" } ] }, { "id": "SDK-rhesis", "name": "Rhesis AI", "vendor": "Rhesis AI", "language": "Python", "type": "Agent测试", "multi_agent": false, "tool_use": true, "memory": false, "streaming": true, "maturity": "早期", "features": [ "Agent测试框架", "8框架对比", "质量评估", "红队测试" ], "pros": [ "测试专精", "框架对比", "质量保障" ], "cons": [ "非Agent框架", "测试工具", "新项目" ], "sources": [ { "source_type": "web", "source_credibility": "B", "collected_at": "2026-06-04T14:03:15.064Z" } ] }, { "id": "SDK-objectbox", "name": "ObjectBox Agent Memory", "vendor": "ObjectBox", "language": "Python/C++", "type": "记忆SDK", "multi_agent": false, "tool_use": true, "memory": true, "streaming": false, "maturity": "生产可用", "features": [ "设备端向量数据库", "嵌入式记忆", "低延迟", "离线运行" ], "pros": [ "设备端运行", "超低延迟", "离线可用" ], "cons": [ "嵌入式场景", "非云端", "规模受限" ], "sources": [ { "source_type": "web", "source_credibility": "B", "collected_at": "2026-06-04T14:03:15.064Z" } ] }, { "id": "SDK-agno", "name": "Agno", "vendor": "社区", "language": "Python", "type": "Agent框架", "multi_agent": true, "tool_use": true, "memory": true, "streaming": true, "maturity": "早期", "features": [ "Agent构建", "工具集成", "记忆管理" ], "pros": [ "简洁", "Python原生", "易上手" ], "cons": [ "新项目", "社区小", "功能有限" ], "sources": [ { "source_type": "web", "source_credibility": "B", "collected_at": "2026-06-04T14:05:21.771Z" } ] }, { "id": "SDK-sim-studio", "name": "Sim Studio", "vendor": "Sim Studio", "language": "TypeScript", "type": "Agent框架", "multi_agent": true, "tool_use": true, "memory": false, "streaming": true, "maturity": "早期", "features": [ "可视化构建", "Agent编排", "工作流设计" ], "pros": [ "可视化", "易用", "工作流" ], "cons": [ "新项目", "功能有限", "社区小" ], "sources": [ { "source_type": "web", "source_credibility": "B", "collected_at": "2026-06-04T14:05:21.771Z" } ] }, { "id": "SDK-genta", "name": "Genta", "vendor": "Genta", "language": "Python", "type": "Agent工具", "multi_agent": false, "tool_use": true, "memory": false, "streaming": true, "maturity": "早期", "features": [ "Agent工具集", "框架比较", "评估工具" ], "pros": [ "工具丰富", "评估功能", "比较功能" ], "cons": [ "非完整框架", "新项目", "社区小" ], "sources": [ { "source_type": "web", "source_credibility": "B", "collected_at": "2026-06-04T14:05:21.771Z" } ] }, { "id": "SDK-alice-labs", "name": "Alice Labs Agent Framework", "vendor": "Alice Labs", "language": "Python", "type": "Agent框架", "multi_agent": true, "tool_use": true, "memory": true, "streaming": true, "maturity": "生产可用", "features": [ "生产测试", "7框架排名", "18+部署经验" ], "pros": [ "生产验证", "丰富经验", "企业级" ], "cons": [ "商业产品", "定制化有限", "成本" ], "sources": [ { "source_type": "web", "source_credibility": "B", "collected_at": "2026-06-04T14:05:21.771Z" } ] }, { "id": "SDK-microsoft-agent-framework", "name": "Microsoft Agent Framework (MAF)", "category": "agent-framework", "description": "Microsoft Agent Framework reached 1.0 GA on April 2, 2026, merging AutoGen and Semantic Kernel into a single supported platform. Supports .NET, Python, and Java.", "features": [ ".NET support", "Python support", "Java support", "AutoGen convergence", "Semantic Kernel convergence", "function calling", "planning", "multi-agent orchestration" ], "status": "active", "year": 2026, "sources": [ "https://devblogs.microsoft.com/agent-framework/microsoft-agent-framework-at-build-2026-announce", "https://learn.microsoft.com/en-us/agent-framework/overview" ] }, { "id": "SDK-morph", "name": "Morph Agent Framework", "category": "agent-framework", "description": "Deep comparison framework covering 8 AI agent SDKs including Claude Agent SDK, OpenAI Agents SDK, Google ADK, LangGraph, CrewAI, Smolagents.", "features": [ "multi-SDK comparison", "agent orchestration", "framework analysis" ], "status": "active", "year": 2026, "sources": [ "https://www.morphllm.com/ai-agent-framework" ] }, { "id": "SDK-rhesis", "name": "Rhesis AI Framework", "category": "agent-framework", "description": "Framework for picking the right agentic framework in 2026. Side-by-side comparison of 8 agent frameworks.", "features": [ "framework comparison", "agent testing", "framework selection" ], "status": "active", "year": 2026, "sources": [ "https://rhesis.ai/post/picking-agentic-framework-2026" ] }, { "id": "SDK-alicelabs", "name": "AliceLabs Agent Framework", "category": "agent-framework", "description": "7 AI agent frameworks ranked by production-readiness: LangGraph #1, Claude Agent SDK #2, CrewAI #3.", "features": [ "production ranking", "agent benchmarking", "framework evaluation" ], "status": "active", "year": 2026, "sources": [ "https://alicelabs.ai/en/insights/best-ai-agent-framework-2026" ] }, { "name": "LangChain / LangGraph", "type": "Framework", "language": [ "Python", "JavaScript", "TypeScript" ], "features": [ "Production-grade AI agent engineering platform", "State management for complex agentic workflows", "Multi-agent orchestration" ], "description": "A leading open-source framework providing tools to build, test, and deploy reliable AI agents with strong support for stateful, multi-step applications.", "status": "Stable / Production-Ready", "sources": [ "[1]", "[2]", "[3]", "[6]", "[7]" ], "id": "SDK-033" }, { "name": "AutoGen / AG2", "type": "Framework", "language": [ "Python" ], "features": [ "Multi-agent conversation patterns", "Group chat capabilities" ], "description": "A Microsoft-backed framework focused on conversational multi-agent interactions, now consolidated under the AG2 branding.", "status": "Stable / Consolidated", "sources": [ "[2]", "[3]", "[5]", "[8]", "[9]" ], "id": "SDK-034" }, { "name": "LangChain", "type": "Framework", "language": "Python, JavaScript/TypeScript", "features": [ "Engineering platform for building AI agents", "Open source frameworks", "Build, test, and deploy reliable AI agents" ], "description": "LangChain provides the engineering platform and open source frameworks developers use to build, test, and deploy reliable AI agents.", "status": "Production", "sources": [ "[1]", "[3]", "[6]", "[7]" ], "id": "SDK-035" }, { "name": "AutoGen (AG2)", "type": "Framework", "language": "Python", "features": [ "Multi-agent conversation framework", "Consolidated major player in 2026" ], "description": "One of the four major players in the 2026 AI agent framework landscape, formerly known as AutoGen and now AG2.", "status": "Production", "sources": [ "[2]", "[3]", "[5]", "[8]" ], "id": "SDK-036" }, { "id": "SDK-037", "name": "Google ADK SDK", "full_name": "@google/adk", "category": "Agent SDK", "description": "Google Agent Development Kit SDK，2025年4月发布，提供层级Agent树构建、多Agent编排和Google Cloud深度集成", "language": "Python", "stars": 15000, "maintainer": "Google", "status": "active", "features": [ "层级Agent树", "Google Cloud集成", "多Agent编排", "开源" ] }, { "id": "SDK-038", "name": "Claude Agent SDK", "full_name": "@anthropic/agent-sdk", "category": "Agent SDK", "description": "Anthropic Claude Agent SDK，2026年6月15日起独立计费，支持文件存储、事件构建和Claude强指令跟随", "language": "Python", "stars": 12000, "maintainer": "Anthropic", "status": "active", "features": [ "Claude原生", "文件存储/加载", "事件构建", "指令跟随" ] }, { "id": "SDK-039", "name": "OpenAI Agents SDK Python", "full_name": "openai-agents", "category": "Agent SDK", "description": "OpenAI Agents SDK，2025年3月发布，轻量Python框架，26900+ Stars，1030万月下载，支持Handoff和工具集成", "language": "Python", "stars": 26900, "maintainer": "OpenAI", "status": "active", "features": [ "轻量框架", "Handoff模式", "工具集成", "1030万月下载" ] }, { "id": "SDK-040", "name": "Smolagents", "full_name": "smolagents", "category": "Agent SDK", "description": "HuggingFace推出的轻量Agent框架，专注简洁和可理解性，适合快速原型和教育场景", "language": "Python", "stars": 7000, "maintainer": "HuggingFace", "status": "active", "features": [ "轻量简洁", "HuggingFace生态", "快速原型", "可理解性" ] }, { "id": "SDK-041", "name": "Microsoft Agent Framework SDK", "full_name": "microsoft-agents", "category": "Agent SDK", "description": "微软统一Agent框架SDK，整合AutoGen/AG2能力，支持Python和C#多语言企业级开发", "language": "Python/C#", "stars": 30000, "maintainer": "Microsoft", "status": "active", "features": [ "AutoGen整合", "多语言", "企业级", "Azure集成" ] } ] } }