For AI Agent Developers
Give your AI agents a real X persona to write from
AI agents are only as good as their context. tweetskill extracts a structured persona from any X account's real tweet history — grounded in engagement data — and exports it in formats your agents can read directly.
The problem
Handcrafted personas drift from reality
Writing a persona prompt by hand is guesswork. It captures what you think the person is like, not what their actual engagement data shows about how they communicate.
No standard format for X personas
Every team invents their own schema for describing a Twitter voice. That makes personas hard to share, version, or use across different agent frameworks.
Personas go stale
Manually updating a persona as someone's posting style evolves is impractical. Fresh data matters.
How tweetskill helps
Data-grounded, not guessed
tweetskill scores every tweet by real engagement — likes, retweets, replies, impressions — and distills the persona from top and bottom performers. The output reflects what actually works, not what seems plausible.
Export in the format your agent needs
Download as skill.md for Claude Code and Hermes, ElizaOS character.json for ElizaOS agents, or raw JSON for anything else. Paste directly into your system prompt or agent config.
Incremental refresh via API
Pro plan includes REST API access and weekly auto-refresh. Keep personas current without manual intervention.
Key features for for ai agent developers
- ✓Export as skill.md — drop into Claude Code or Hermes SKILL.md
- ✓Export as ElizaOS character.json — ready for ElizaOS agents
- ✓Export as raw JSON — use in any agent or system prompt
- ✓Engagement-scored persona extraction — grounded in real data
- ✓Public persona gallery — browse and use community-built personas
- ✓REST API for programmatic access (Pro)
Personas worth studying
Public profiles you can analyze and generate content from right now.
@elonmusk
High-contrast persona — ultra-short amplifiers, institutional takedowns, deadpan product drops. A good stress test for voice fidelity.
@VitalikButerin
Technical depth + accessibility. Useful for agents that need to explain complex topics clearly.
@naval
Dense aphoristic style — tests whether your agent can maintain a consistent philosophical voice.
Free plan available. 7-day free trial on paid plans. No credit card required.