binary-knight/logwhisperer — reverse-engineered prompt

Reverse engineered prompt

Build me a self hosted tool called LogWhisperer for Linux servers that can read logs and explain what is going on in plain English using a local Ollama model. I want it to summarize recent problems from system logs, regular log files, and Docker containers, and help spot patterns instead of just dumping raw lines. It should feel production ready, with deduping, caching, and sensible rate limiting so repeated noise does not spam me.

I also want a monitoring mode that keeps watching logs in real time and sends Discord webhook alerts for serious issues, including optional user or role mentions for error levels. Please make it easy to run either from source or with Docker Compose, and include a simple config file example for model settings, Ollama host, and alert rules. If there is already a web or API piece in the repo, wire it up so it works cleanly too. Keep the setup straightforward for a normal server admin, and check current docs online if you need to fill in anything.

Want more depth? Deep Reverse