@modusop/mcp-server

@modusop/mcp-server

Acts as a stdio-to-HTTP proxy for Modus Brain, enabling MCP-compatible AI clients to access an organization's knowledge base in ModusOp.

Category
访问服务器

README

@modusop/mcp-server

Stdio→HTTP proxy for Modus Brain — gives any MCP-aware AI client (Claude Desktop, Cursor, Cline, Continue.dev, etc.) access to your ModusOp organisation's knowledge base.

What it does

This package is a thin proxy. It speaks the MCP stdio transport on one end and forwards every JSON-RPC request to Modus Brain's HTTP MCP endpoint on the other.

All tool implementations live in Modus Brain itself — when ModusOp ships a new Brain tool, every install of this package picks it up automatically on the next request. No re-publish, no version bump, no user action needed.

If your AI client speaks remote HTTP MCP natively, you don't need this package at all — point it directly at https://brain.modusop.app/mcp. This package exists for clients that only speak stdio.

Setup

1. Get a Brain token

Sign in at brain.modusop.app, generate a token labelled for the device or person who'll use it, copy the plaintext value (you only see it once).

2. Add to your AI client's MCP config

Claude Desktop — ~/Library/Application Support/Claude/claude_desktop_config.json (macOS) or %APPDATA%\Claude\claude_desktop_config.json (Windows):

{
  "mcpServers": {
    "modusop-brain": {
      "command": "npx",
      "args": ["-y", "@modusop/mcp-server"],
      "env": {
        "MODUSOP_API_TOKEN": "mo_brain_..."
      }
    }
  }
}

Cursor / Cline / Continue.dev — .cursor/mcp.json (or the equivalent):

{
  "mcpServers": {
    "modusop-brain": {
      "command": "npx",
      "args": ["-y", "@modusop/mcp-server"],
      "env": {
        "MODUSOP_API_TOKEN": "mo_brain_..."
      }
    }
  }
}

3. Restart your client

Quit and reopen. The Modus Brain tools should appear in the MCP indicator within a few seconds.

Env vars

Variable Required Default Notes
MODUSOP_API_TOKEN yes — Brain token from brain.modusop.app
MODUSOP_BRAIN_URL no https://brain.modusop.app/mcp Override only for local Brain development

Available tools

Whatever Brain exposes — tools/list reflects the current set. As of v1.0.0:

  • whoami — identify the user + org this token is scoped to
  • search_knowledge — semantic search over the org KB
  • get_chunk — deep-dive on a specific search result
  • list_recent — recently indexed KB items
  • get_client_brief — synthesised client one-pager
  • add_observation — append a note
  • add_decision — record context + decision + rationale
  • link_observation — typed link between two chunks
  • request_delete — admin-gated removal

Migrating from 0.x

Versions 0.x of this package implemented tools in-package and talked to ModusOp's older /api/mcp endpoint with an MO API token from /settings/api-tokens. v1.0.0 retires that path:

  • Generate a new Brain token at brain.modusop.app, not an MO API token
  • Replace the value of MODUSOP_API_TOKEN with the new token
  • No other config change needed — same command, same args, same env var name

Old tools (search_projects, search_clients, get_client_context, start_timer, etc.) are no longer present in this package's catalogue. They've been retired in favour of Brain's tools, which cover the same surface plus knowledge-base search and append-writes. If you need the old tool set, pin to @modusop/mcp-server@0.2.5 — it still works but won't receive updates.

Local development

git clone https://github.com/jwwd-repos/modusop-mcp.git
cd modusop-mcp
npm install
MODUSOP_API_TOKEN=... npm run dev

The proxy reads JSON-RPC on stdin and writes responses on stdout, so test it manually with:

echo '{"jsonrpc":"2.0","id":1,"method":"tools/call","params":{"name":"whoami"}}' \
  | MODUSOP_API_TOKEN=... npm run dev

Licence

See LICENSE.md.

推荐服务器

Baidu Map

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。

官方
精选
JavaScript
Playwright MCP Server

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。

官方
精选
TypeScript
Magic Component Platform (MCP)

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。

官方
精选
本地
TypeScript
Audiense Insights MCP Server

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。

官方
精选
本地
TypeScript
VeyraX

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。

官方
精选
本地
graphlit-mcp-server

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。

官方
精选
TypeScript
Kagi MCP Server

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。

官方
精选
Python
e2b-mcp-server

e2b-mcp-server

使用 MCP 通过 e2b 运行代码。

官方
精选
Neon MCP Server

Neon MCP Server

用于与 Neon 管理 API 和数据库交互的 MCP 服务器

官方
精选
Exa MCP Server

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。

官方
精选