About LLM Atlas

Good prompts are scattered across Reddit threads, gists and blog posts — with no way to tell which ones actually work. LLM Atlas is a free, open map of prompts with a quality signal you can trust.

What makes a prompt trustworthy here

• Tested-on badges. Authors mark the exact model versions they ran a prompt on.
• Example outputs. Prompts can attach the real output they produced — proof, not promises.
• Stars, not ratings soup. One GitHub-style signal: star what's good.

Free means free

Browsing, copying, starring and publishing are free, and always will be. Prompts are published under open licenses (CC0 by default), so you can use them anywhere — including commercially.

Model-aware by design

A prompt that sings on Claude can stumble on GPT, and a Midjourney style recipe means nothing to Veo. Every prompt on the atlas is tagged by model family and version, so you can browse by the model you actually use.

Join the survey

Found a prompt that works? Chart it. Tested someone else's on a new model? Say so in the comments. The map improves with every expedition.