xkonjin/tweet-smash-mcp-server

TweetSmash MCP Server

A Node.js Express server that bridges ChatGPT Custom Connector with TweetSmash API using Server-Sent Events.

Requirements

• Node 18+ (uses the built-in global fetch)

Local Development

# 1. Install dependencies
npm install

# 2. Create a .env file and set your TweetSmash API key
#    (or export it in your shell)
TWEETSMASH_API_KEY=xxxxxxxx npm start

The server starts on http://localhost:3000. Endpoints:

• POST /invoke
• GET /health

Deploying to Vercel

1. Install the Vercel CLI (optional but handy):

npm i -g vercel

2. Link your project and create a new Vercel deployment:

vercel # follow prompts

3. Add the required environment variable in the Vercel dashboard or via CLI:

vercel env add TWEETSMASH_API_KEY

4. Redeploy (if you added the env var via the dashboard):

vercel deploy --prod

After deployment your endpoints will be available at:

https://<your-project>.vercel.app/invoke
https://<your-project>.vercel.app/health

ChatGPT Custom Connector Configuration

1. MCP Server URL: https://<your-project>.vercel.app/invoke
2. Authentication: Bearer Token (value: your TweetSmash API key, optional — if provided in the request it overrides the environment variable)

API Details

POST /invoke

• Body: { "tool_input": { "query": "<optional keyword>" } }
• Response: Server-Sent Events (single event) containing TweetSmash bookmark JSON.
• The server always fetches up to 100 bookmarks, then filters them locally (case-insensitive match in tweet text or tags).
  • You can append a 4-digit year (e.g. "defi 2024") to restrict results to that year.

GET /health

• Simple health-check JSON.