llm-council

Why we included this project

Ever wanted a second opinion on a hard question without juggling multiple chat windows? This project shows one way to get it: a local web app that sends your prompt to several LLMs through OpenRouter, shows each answer in tabs, then has the models anonymously review and rank one another before a designated Chairman compiles a single final response. You get both the synthesized answer and the raw individual takes, which makes it handy for seeing which model actually handles a given kind of question best. It's also a straightforward example of multi-model orchestration that's easy to run locally and tweak. Just know the author calls it an unsupported Saturday hack for inspiration, not a maintained product, so treat it as a starting point to adapt rather than something to build on.