I reverse engineered @cluely – and their desktop source code exposes their entire system prompts and models used. What's inside? 🧵
Cluely instructs the prompt to "NEVER mention the specific LLM providers", but we can see that Cluely uses GPT 4.1 (in most cases) and can also use Claude Sonnet 3.7.
The app also includes Cluely's 400+ line enterprise prompt – featuring examples about dolphins, conversations about Databricks, and chats with Palantir interns.
Check out the full prompts:
1,16M