How contributing works

1. Add a conversation: enter a prompt and response, paste a transcript, or upload a ChatGPT or Claude export.
2. We parse it and scan it for sensitive information.
3. You review every message, fix speakers, and redact or delete anything you don’t want published.
4. You confirm each consent statement (version 2026-11-15):
   • I want this conversation published publicly on The Chat Archive, where anyone can read it.
   • I understand it may be downloaded and redistributed in open research datasets under CC BY 4.0, and that removing it later cannot recall copies that were already downloaded.
   • This is a real conversation with an AI system, and I have the right to share it.
   • It does not contain identifiable private information about other people without their permission.
   • I reviewed the entire conversation and redacted anything I do not want published.
5. We check for exact duplicates and suppressed content, and flag likely near-duplicates.
6. A moderator reviews it for privacy, authenticity and quality.
7. If approved, it’s published and earns Research Points.

What makes a good contribution

• Real conversations, copied faithfully. Don’t edit the AI’s words except to redact.
• An accurate model name if you know it. It’s shown as self-reported.
• Interesting failures, corrections and evaluations, and topics we have little of.

Fabricated conversations are rejected, and confirmed fraud reverses Research Points.