Twitter Monitor fetches your X/Twitter home timeline using the internal GraphQL API (no official API key needed), clusters tweets by topic using AI embeddings, and generates narrative digests with cited sources.

Workflow

1. Fetch — Pull tweets from your timeline via X's internal API
2. Embed — Convert tweets to vectors using sentence-transformers
3. Cluster — Group related tweets by semantic similarity (K-means or HDBSCAN)
4. Summarize — Generate narrative digests with Claude, citing original tweets

Features

• No API Key Required — Uses X's internal GraphQL API with browser session tokens
• Semantic Clustering — Groups tweets by meaning, not just keywords
• Verified Citations — Every claim links back to the source tweet
• Parallel Processing — Summarizes clusters concurrently for speed

Requirements

• Python 3.10+
• X/Twitter account (for auth tokens extracted from browser cookies)
• Claude Code or Claude API (for summarization step)