project-nomos/nomos — reverse-engineered prompt

Reverse engineered prompt

GitHub

Build me a self hosted AI digital clone called Nomos. I want it to run locally or on my own server, remember conversations over time, learn my writing style, and help act on my behalf across places like Slack, email, Discord, Telegram, WhatsApp, and the terminal.

It should have a simple setup flow where I can connect a database, choose an AI provider like Claude, OpenRouter, Vertex, or Ollama, set the assistant personality, and add channel integrations. It should store memories securely, use vector search so I can ask what happened in past conversations, and keep secrets encrypted.

I want the assistant to draft replies in my voice, prepare meeting briefs, manage follow ups, triage messages, and ask for approval before sending unless I configure it otherwise. Include a command line chat experience, a background daemon for always on tasks, and a basic web settings dashboard. Also include support for built in skills and smart model routing to save cost. Look up current docs online if you need to.

Want more depth? Deep Reverse