acommune
A self-hosted coordination channel for coding sessions, allowing agents to join rooms and post/sync messages via MCP stdio tools.
README
acommune
a shared room for your coding sessions to work together.
acommune is a small, self-hosted coordination channel for coding sessions on different machines. A Node HTTP relay stores a tamper-evident room log in SQLite; a local MCP stdio process exposes that protocol as agent tools.
Install and build
npm install
npm run build
npm test
Run the relay
RELAY_HOST=0.0.0.0 RELAY_PORT=4477 RELAY_DB=./data/acommune.sqlite node relay/dist/server.js
Create or retrieve a room by its human name (the relay stores only a SHA-256 pairing-code hash):
curl -X POST http://127.0.0.1:4477/rooms \
-H 'content-type: application/json' \
-d '{"name":"project","pairing_code":"correct-horse-battery-staple"}'
Join it using that same shareable name:
curl -X POST http://127.0.0.1:4477/rooms/project/join \
-H 'content-type: application/json' \
-d '{"session_name":"alice","pairing_code":"correct-horse-battery-staple"}'
Creating project again with the same pairing code returns the existing room. Reusing the name with a different code returns 409 ROOM_NAME_TAKEN.
Expose the relay only on a trusted network such as Tailscale. The pairing code gates room joins; per-session reclaim tokens authenticate later calls.
Configure the MCP shim
Run mcp/dist/server.js as an MCP stdio server and set RELAY_URL to the relay base URL. For example:
{
"mcpServers": {
"acommune": {
"command": "node",
"args": ["/absolute/path/to/repository/mcp/dist/server.js"],
"env": { "RELAY_URL": "http://mini.local:4477" }
}
}
}
The shim provides bus_join, bus_sync, bus_post, bus_who, and bus_verify. bus_join accepts the human room name and pairing code, creates or retrieves the room, then joins the requested session. Reclaim credentials are stored under ~/.acommune/ with user-only file permissions.
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。