Back to Skills Hub
LLM Council

LLM Council

@am-will
developmentMulti-Agent OrchestrationBias-Resistant PlanningParallel LLM Execution

A multi-agent orchestration system that launches multiple AI planners in parallel, collects their independent plans, anonymizes them to reduce bias, and uses a judge agent to evaluate and merge the best elements into a high-quality final implementation plan.

🚀 LLM Council orchestrates multiple AI agents in parallel to create superior implementation plans. It launches different AI models simultaneously, collects their independent proposals, anonymizes them to eliminate bias, and uses a judge agent to evaluate and merge the best ideas into one final, high-quality plan.

💡 Perfect for complex projects where you need diverse perspectives and robust solutions. Use it to generate technical strategies, business plans, or detailed implementation roadmaps while reducing blind spots and single-perspective limitations.

✨ Unlike single-agent approaches, LLM Council combines the strengths of multiple AI models, automatically retries failures, and provides real-time visualization so you can watch, compare, and refine results iteratively.

GitHub

Requirements

OpenAI Codex

OpenAI Codex API for code generation planning

Anthropic Claude

Anthropic Claude API for planning generation

Google Gemini

Google Gemini API for planning generation

Python 3.10+

Python runtime environment