LemniscateIn development
An LLM that stays in the loop and keeps improving a codebase. Connect a GitHub, GitLab or GitVerse token, point the agent at any OpenAI-compatible endpoint (OpenAI, Azure, vLLM, Ollama, LM Studio), and on a schedule you set it analyses your repositories, proposes fixes and improvements, writes the code, pushes a branch and opens pull requests.
Highlights
- — Works with any OpenAI-compatible LLM endpoint, using your own keys and models
- — Runs the agent loop as background jobs in a separate Redis + BullMQ worker, so long analyses stay off the request path
- — Delivers proposals as branches and pull requests on your own git host, with an optional second-pass LLM review and automatic merge
- — Streams the agent thinking, editing, committing and pushing over SSE as it happens
Stack
TypeScript · Fastify · React · BullMQ · PostgreSQL · Docker · LLM