Vitalcheffe/iris-gateway — reverse-engineered prompt
Reverse engineered prompt
Build me a self hosted personal AI messaging gateway called Iris. I want to run it on my own server and talk to one assistant from Telegram, WhatsApp, Discord, Slack, and a simple web chat, with all conversations sharing the same memory in a local SQLite vault. It should use OpenRouter free models and have an easy setup wizard that asks for API keys, bot tokens, model choice, enabled channels, and writes a config file.
The assistant should remember my goals, infer basics like language and timezone from chats, follow up about deadlines, and support useful tools like calendar, email, contacts, tasks, web search, and vault search. Add sensible security so only allowed users and channels can use it, with rate limits and policy settings for tool access. Include a health endpoint, clear logs, Docker support, and CLI commands to init, run, and check status. Use TypeScript and Node, keep it production ready, and look up current docs online if you need to.
Want more depth? Deep Reverse