LLM Council
/About

Company

About LLM Council

LLM Council is a multi-model AI deliberation platform. Instead of asking one AI and taking its answer at face value, LLM Council puts your question to a panel of models — each thinking independently, then reviewing each other's work, before a chairman synthesises a final answer.

The idea is inspired by karpathy/llm-council— Andrej Karpathy's experiment in multi-model deliberation. We built a hosted, production-ready version with a clean interface, bring-your-own-key support, and streaming real-time updates.

The platform supports 7 LLM gateways — from hosted services like OpenRouter, Portkey, and Helicone, to self-hosted stacks like LiteLLM, Bifrost, KrakenD, and Cloudflare AI Gateway.

The three stages

Stage 1
Independent Opinions

Each model answers the question without seeing others' responses.

Stage 2
Blind Peer Review

Each model critiques and ranks the others' answers anonymously.

Stage 3
Chairman Synthesis

One model reads everything and writes the definitive answer.

Contact us