WilmerAI

Why we included this project

WilmerAI sits between your application and the LLM APIs behind it, and it routes each request by looking at the whole conversation, not just the latest message. That context matters: a question like "What do you think it means?" gets treated as a follow-up to whatever you were just discussing, rather than a fresh prompt. The routing logic itself is a node-based workflow defined in JSON, so each step can call a different model, run a tool or script, or hand off to another workflow, and the whole chain appears to your front end as one ordinary API call. That makes it a good fit when you want to mix local models with cloud APIs and control exactly how prompts are rewritten along the way. It is still under active development, so treat it as a promising tool rather than a finished product.