clawd-mcp
Bridges Cursor and Claude Desktop with the OpenClaw and Moltbook ecosystems to manage personal AI agents and communication gateways across platforms like WhatsApp and Discord. It provides MCP tools to invoke agents, manage sessions, and handle social agent interactions within a secure, virtualized environment.
README
openclaw-mcp
MCP server + webapp that bridge Cursor and Claude Desktop to OpenClaw and Moltbook. openclaw-mcp uses those platforms (Gateway, APIs); it does not implement or replace them. Alpha. Ingest for LLMs: gitingest.com/sandraschi/openclaw-mcp.
What this repo is
| Part | What it does |
|---|---|
| MCP server (stdio) | FastMCP 2.14+ tools: agent, sessions, channels, routing, skills, gateway, security, moltbook. For Cursor/Claude Desktop. |
| Webapp (React + Vite + Tailwind) | Dashboard on port 5180: Startpage, AI (Ollama), Channels, Routes, Diagram, Statistics, Moltbook, Integrations, Clawnews, Skills, Security, Generate landing (landing-site generator), Settings. |
| webapp_api (FastAPI) | Backend on 5181: /api/ask, /api/gateway/status, /api/skills, /api/clawnews, /api/ollama/*, /api/channels, /api/routing, /api/openclaw/status, /api/landing-page. |
One place to run agents, manage channels/routes/skills, and use Moltbook; OpenClaw and Moltbook stay separate.
Install & run
Clone-based (no PyPI package yet). See INSTALL.md for: clone repo, install from source, MCP only or webapp (API + frontend), one-shot scripts, config, logging, checks.
Repo layout
- src/openclaw_mcp/ – MCP server and tools
- webapp/ – React dashboard (port 5180)
- webapp_api/ – FastAPI backend (port 5181)
- scripts/ – install.ps1, install.bat, start.ps1, start.bat, check.ps1, serve_logs.ps1, mcpb-build.ps1
- snippets/ – MCP config snippet (snippets/mcp-config-openclaw-mcp.json); see INSTALL.md and mcp-central-docs pattern
Architecture – data flow, MCP vs API.
Docs
| Doc | Description |
|---|---|
| LICENSE | MIT license |
| CONTRIBUTING.md | How to contribute |
| INSTALL.md | Install, run, config, checks |
| docs/README_INDEX.md | Doc index |
| docs/README_WEBAPP.md | Webapp pages, API, Logger |
| docs/README_OPENCLAW_MCP_TOOLS.md | MCP tools |
| docs/README_OPENCLAW.md | OpenClaw (external) |
| docs/README_MOLTBOOK.md | Moltbook (external) |
| docs/HOW_THIS_WAS_MADE.md | How this was made (vibecode, Cursor, one day) |
| SECURITY.md | Threats, hardening |
Security
OpenClaw has major security risks. Use clawd_security; prefer VM and loopback. SECURITY.md, docs/SECURITY_HARDENING.md.
Repo manifests (LLM scrapers)
| File | Triggers | Purpose |
|---|---|---|
| glama.json | Glama GitHub scraper | MCP server listing: claim ownership, metadata, Docker, usage. Requires $schema + maintainers (GitHub usernames). Re-run claim flow on Glama after changes. |
| llms.txt | Gitingest, llmstxt.org | LLM-friendly manifest: H1 + blockquote summary + ## sections with links. Improves repo ingestion for LLMs (e.g. gitingest.com/sandraschi/openclaw-mcp). |
Other scrapers: no extra files needed. Gitingest ingests the repo (replace github.com with gitingest.com in the repo URL); llms.txt gives it a curated entry point. Glama is the main MCP-directory scraper that uses glama.json; others (e.g. Cursor’s MCP discovery) may crawl GitHub without a manifest.
References
This README is formatted to render well on the GitHub repo page (pretty-printed).
推荐服务器
Baidu Map
百度地图核心API现已全面兼容MCP协议,是国内首家兼容MCP协议的地图服务商。
Playwright MCP Server
一个模型上下文协议服务器,它使大型语言模型能够通过结构化的可访问性快照与网页进行交互,而无需视觉模型或屏幕截图。
Magic Component Platform (MCP)
一个由人工智能驱动的工具,可以从自然语言描述生成现代化的用户界面组件,并与流行的集成开发环境(IDE)集成,从而简化用户界面开发流程。
Audiense Insights MCP Server
通过模型上下文协议启用与 Audiense Insights 账户的交互,从而促进营销洞察和受众数据的提取和分析,包括人口统计信息、行为和影响者互动。
VeyraX
一个单一的 MCP 工具,连接你所有喜爱的工具:Gmail、日历以及其他 40 多个工具。
graphlit-mcp-server
模型上下文协议 (MCP) 服务器实现了 MCP 客户端与 Graphlit 服务之间的集成。 除了网络爬取之外,还可以将任何内容(从 Slack 到 Gmail 再到播客订阅源)导入到 Graphlit 项目中,然后从 MCP 客户端检索相关内容。
Kagi MCP Server
一个 MCP 服务器,集成了 Kagi 搜索功能和 Claude AI,使 Claude 能够在回答需要最新信息的问题时执行实时网络搜索。
e2b-mcp-server
使用 MCP 通过 e2b 运行代码。
Neon MCP Server
用于与 Neon 管理 API 和数据库交互的 MCP 服务器
Exa MCP Server
模型上下文协议(MCP)服务器允许像 Claude 这样的 AI 助手使用 Exa AI 搜索 API 进行网络搜索。这种设置允许 AI 模型以安全和受控的方式获取实时的网络信息。