twitterapi-io-mcp
Offline TwitterAPI.io documentation for Claude and AI assistants via MCP. 76 active endpoints (with full method/params/body), 32 guides, 24 blog posts. No API key needed.
[npm version]
[npm downloads]
[Smithery]
[GitHub stars]
[License: MIT]
What is this?
An MCP server that gives Claude, Cursor, VS Code Copilot, and other AI assistants instant, offline access to TwitterAPI.io documentation.
- 76 active API endpoints with full HTTP method, body/query parameters, cURL examples, and response schemas
- 32 guide pages covering pricing, authentication, rate limits, webhooks, streaming
- 24 blog posts with tutorials and use cases
- All v2 write endpoints documented with
login_cookies, proxy, and body params
- All deprecated v1 endpoints removed — only current v2 endpoints remain
- Zero API key required — everything is bundled locally
Unlike other Twitter MCP servers that proxy live API calls (and need your API key), this one ships a complete documentation snapshot. Your AI assistant reads it locally, instantly.
Recent Changes
- v1.2.0: OpenAPI-first scraper rewrite — structured parameters with types, response schemas, eliminates Vercel checkpoint issues
- v1.1.x: Added
list_timeline + get_user_timeline, removed 7 deprecated V1 endpoints, security fixes (hono, ajv, express-rate-limit)
- v1.0.23: Platform advisory — Twitter disabled
since:/until: search operators. Use since_time:UNIX / until_time:UNIX instead
- v1.0.22: Added 7 endpoints (
get_space_detail, get_tweet_replies_v2, get_user_about, etc.)
See CHANGELOG.md for full history.
Disclaimer: Independent community project. Not affiliated with TwitterAPI.io.
Installation
Previously published as twitterapi-docs-mcp? Just change the package name — everything else stays the same.
Requires: Node.js 18.18.0+
Claude Code (recommended)
Claude Desktop
Add to your config file:
- macOS:
~/Library/Application Support/Claude/claude_desktop_config.json
- Windows:
%APPDATA%\Claude\claude_desktop_config.json
- Linux:
~/.config/Claude/claude_desktop_config.json
Cursor
Add to .cursor/mcp.json in your project root:
VS Code / VS Code Insiders
Add to .vscode/mcp.json in your project root:
Windsurf
Add to ~/.codeium/windsurf/mcp_config.json:
Docker
Or build from source:
Available tools
For AI assistants
If you're an AI assistant using this MCP server, here's a quick reference:
Tips:
- Search is fuzzy and typo-tolerant:
"twet object" still finds results
- Use
max_results: 5 for focused results
- All endpoints are current v2 — no deprecated endpoints remain
What can you ask?
Here are real prompts that work well with this MCP server:
- "What are the rate limits for TwitterAPI.io?" — triggers
get_twitterapi_guide
- "Show me the tweet advanced search endpoint" — triggers
get_twitterapi_endpoint
- "How do I authenticate with the API?" — triggers
get_twitterapi_auth
- "What's the pricing?" — triggers
get_twitterapi_pricing
- "List all user-related endpoints" — triggers
list_twitterapi_endpoints with category: "user"
- "How do webhook filter rules work?" — triggers
search_twitterapi_docs
- "What write endpoints need login?" — triggers
search_twitterapi_docs with query: "login_cookies"
- "How do I upload media and create a tweet?" — triggers sequential
get_twitterapi_endpoint calls
All 76 endpoints by category
Configuration
Troubleshooting
npx hangs or fails on first run
The first run downloads the package from npm. If it hangs, try:
Or install globally: npm install -g twitterapi-io-mcp
"Node.js version not supported"
This server requires Node.js 18.18.0 or later. Check your version:
claude mcp add not recognized
Make sure Claude Code CLI is installed and up to date:
Server starts but no tools appear
Restart your MCP client after adding the server config. Most clients (Claude Desktop, Cursor) need a restart to detect new MCP servers.
How it works
- AI assistant calls an MCP tool (search, get endpoint, etc.)
- Server searches the bundled documentation snapshot
- Results return instantly with no network requests
Why I built this
I use TwitterAPI.io daily for my projects. Switching between my editor and their docs site got old fast. So I packaged the entire documentation into an MCP server. Now I ask Claude and get answers without leaving my terminal.
If you work with the Twitter/X API through TwitterAPI.io, this saves you the same context-switching. Install it, forget about it, and just ask.
Development
Not using MCP?
If your AI tool doesn't support MCP (like OpenClaw), use the skill version instead — a single markdown file with all 76 active endpoints:
👉 twitterapi-io-skill — works with OpenClaw, Claude, GPT, Gemini, and any LLM.
Links
Contributing
PRs welcome! Fork the repo, create a branch, make your changes, and open a pull request. Bug reports and feature requests go to GitHub Issues.
License
MIT
Acknowledgments