Limitless MCP Server

Limitless MCP Server

Enables Claude to search, retrieve, and analyze Lifelogs from Limitless AI through a Cloudflare Workers deployment with GitHub OAuth authentication.

Category
访问服务器

README

Limitless MCP Server for Cloudflare Workers

A Model Context Protocol (MCP) server that provides Claude with access to your Limitless AI lifelogs. Deploy to Cloudflare Workers with GitHub OAuth authentication for secure access.

About Limitless

Limitless is a personalized AI wearable device (the Pendant) that helps you capture and recall conversations, meetings, and thoughts throughout your day. The Pendant features:

  • 🎙️ Conversation Recording - Captures meetings, personal reflections, and spontaneous thoughts with crystal clear audio
  • 🔖 Moment Bookmarking - Tap the device to bookmark important moments
  • 🧠 AI-Powered Insights - Ask questions about your recorded conversations and get personalized insights
  • 🔒 Privacy-First - HIPAA-compliant with permission-based data protection
  • 🔋 All-Day Battery - Lightweight wearable with durable aluminum body

This MCP server allows Claude to access your Limitless lifelogs, enabling you to search, retrieve, and analyze your recorded conversations directly within Claude.

Note: This project requires manual setup rather than one-click deployment because it needs secure API keys (Limitless AI, GitHub OAuth) and creates KV namespaces for OAuth token storage. The setup process ensures your credentials remain private and secure.

Features

  • 🔍 Search Lifelogs - Hybrid semantic and keyword search across all your recordings
  • 📝 Retrieve Full Content - Get complete lifelog content including transcripts and summaries
  • 📅 List Recent Recordings - Browse your latest lifelogs with date filtering
  • 🔐 Secure Authentication - GitHub OAuth integration for secure access
  • ⚡ Edge Deployment - Runs on Cloudflare Workers for low latency
  • 🎯 Claude Integration - Seamlessly works with Claude.ai as a custom connector

Prerequisites

Cloudflare Workers Free Tier

This project runs perfectly on Cloudflare's free tier, which includes:

  • 100,000 requests per day (more than enough for personal use)
  • 10ms CPU time per invocation (sufficient for API calls)
  • Up to 1MB KV storage (plenty for OAuth tokens)
  • Unlimited Durable Objects (for MCP agent state)
  • No credit card required

For typical usage with Claude, you'll likely use less than 1% of these limits.

Quick Start

1. Clone and Install

git clone https://github.com/BurtTheCoder/mcp-limitless.git
cd mcp-limitless
npm install

2. Initial Setup

Run the automated setup script:

./setup.sh

This will:

  • Create your wrangler.toml configuration
  • Set up KV namespaces for OAuth storage
  • Create a .dev.vars template for local development

3. Configure API Keys

Get your Limitless API Key

  1. Go to app.limitless.ai
  2. Navigate to Settings → API
  3. Generate or copy your API key

Create GitHub OAuth App

  1. Go to GitHub Settings → Developer settings → OAuth Apps
  2. Click "New OAuth App"
  3. Fill in:
    • Application name: Limitless MCP Server
    • Homepage URL: https://your-worker.workers.dev (you'll get this after deploy)
    • Authorization callback URL: https://your-worker.workers.dev/github/callback
  4. Save your Client ID and Client Secret

Update Local Configuration

Edit .dev.vars:

LIMITLESS_API_KEY="your-limitless-api-key-here"
GITHUB_CLIENT_ID="your-github-client-id"
GITHUB_CLIENT_SECRET="your-github-client-secret"

4. Test Locally

npm run dev

Your server will be available at http://localhost:8787

Test the OAuth flow by visiting: http://localhost:8787/authorize

5. Deploy to Cloudflare

Run the deployment script:

./deploy.sh

This will prompt you to enter your secrets securely and deploy to Cloudflare Workers.

You'll get a URL like: https://limitless-mcp-server.your-account.workers.dev

6. Update GitHub OAuth App

After deployment, update your GitHub OAuth App's callback URL to:

https://limitless-mcp-server.your-account.workers.dev/github/callback

7. Connect to Claude

  1. Go to Claude.ai
  2. Navigate to Settings → Connectors
  3. Click "Add custom connector"
  4. Enter your Workers URL with /sse endpoint:
    https://limitless-mcp-server.your-account.workers.dev/sse
    
  5. Click "Add" and authorize with GitHub when prompted
  6. Enable the connector in your chat

Available Tools in Claude

Once connected, Claude will have access to these tools:

  • Limitless Ai:search_lifelogs - Search using natural language or boolean operators

    • Example: "meetings about the new project"
    • Example: "dinner OR lunch with Bob"
  • Limitless Ai:get_lifelog - Retrieve a specific lifelog by ID

    • Returns full content including transcript and summaries
  • Limitless Ai:list_recent_lifelogs - List your recent recordings

    • Supports date filtering and pagination
    • Can filter by starred status

Project Structure

mcp-limitless/
├── src/
│   ├── github-oauth-index.ts  # Main entry point with OAuth
│   ├── mcp-agent.ts           # MCP server implementation
│   ├── simple-index.ts        # Alternative without GitHub OAuth
│   └── index.ts               # Alternative implementation
├── wrangler.toml.example      # Cloudflare config template
├── setup.sh                   # Automated setup script
├── deploy.sh                  # Deployment script
├── package.json               # Dependencies
└── README.md                  # This file

Configuration Options

Environment Variables

Variable Required Description
LIMITLESS_API_KEY Yes Your Limitless AI API key
GITHUB_CLIENT_ID Yes GitHub OAuth App Client ID
GITHUB_CLIENT_SECRET Yes GitHub OAuth App Client Secret
ENABLE_IP_ALLOWLIST No Set to "true" to only allow Anthropic IPs
RATE_LIMIT_KV No KV namespace binding for rate limiting

Customization

To use without GitHub OAuth, change the main entry in wrangler.toml:

main = "src/simple-index.ts"  # Simple approval page instead of GitHub

Troubleshooting

Connection Issues

  1. "MCP error -32000: Connection closed"

    • Ensure you're using the /sse endpoint
    • Check that OAuth authentication completed successfully
  2. "Tool not found" errors in Claude

    • Tools are namespaced as Limitless Ai:tool_name
    • Claude should automatically detect this
  3. 404 errors from Limitless API

    • Verify your API key is correct
    • Check that your Limitless account has API access enabled

Viewing Logs

# Stream live logs from your worker
npx wrangler tail

# Or view in Cloudflare Dashboard
# Go to Workers & Pages → your-worker → Logs

Rate Limits

  • Limitless API: 180 requests per minute
  • Cloudflare Workers Free: 100,000 requests per day

Security Features

Built-in Security

  • OAuth 2.0 with PKCE for secure authorization
  • API keys stored as encrypted Cloudflare secrets
  • OAuth tokens expire after 1 hour
  • GitHub OAuth provides identity verification
  • HTTPS enforced by Cloudflare Workers

Optional Enhanced Security

  • IP Allowlisting: Restrict access to Anthropic's official IPs only
  • Rate Limiting: Prevent abuse with configurable request limits
  • Security Headers: HSTS, CSP, X-Frame-Options automatically applied

To enable enhanced security features, see SECURITY.md for detailed configuration.

Contributing

Contributions are welcome! Please:

  1. Fork the repository
  2. Create a feature branch
  3. Make your changes
  4. Test locally and with deployment
  5. Submit a pull request

License

MIT License - See LICENSE file for details

Support

Acknowledgments

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