A curated index of projects in the LLM Council ecosystem — multi-model deliberation systems where multiple LLMs collaborate, debate, and synthesize answers. Inspired by Andrej Karpathy's original llm-council.
Disclaimer: This is a point-in-time snapshot (April 2026), non-exhaustive, and does not constitute endorsement of any listed project. Some projects appear in more than one section where they fit multiple buckets.
Religious Values Benchmark Meta-Evaluation Dashboard — 401 candidates, 14 traditions, 9 LLMs, 12-member expert council review.
Language: HTML
Author: cjimmylin
Self-improvement loops for OpenClaw agents via multi-model LLM council.
Language: Python
Author: shadmau
Multi-agent orchestration skill for OpenClaw — parallel deliberation across specialized AI personas with auditable transcripts.
Language: Python
Author: infektyd
Enhance OpenClaw with council mode, running multiple role-based prompts for cross-checked, transparent final answers and risk analysis.
Language: Python
Author: nonaammme
4 Councilmen Model (4CM): Non-convergent Multi-agent Coordination.
Language: Python
Author: Klastrovanie
AI Council is a blazing-fast, open-source platform for running structured multi-LLM debates. Build a diverse cast of AI personas, back them with different state-of-the-art models, inject unique system prompts, and watch them argue, challenge each other, and converge on a verdict — all streaming live in your browser.
Language: TypeScript
Author: meisamsharahi
Multi-LLM Council Interface — query multiple models simultaneously, compare responses side-by-side, and evaluate with structured criteria. A Sanctum partner tool.
Language: TypeScript
Author: TheAIHorizon
Karpathy-style LLM Council using only GitHub Copilot OAuth. 3x3 multi-model deliberation with anonymized peer review and chairman synthesis.
Language: Python
Author: hellofrommorgan
A council of free OpenRouter LLMs deliberating your question. Just for fun.
Language: TypeScript
Author: flickmediasa
Multi-perspective LLM deliberation CLI. Expert personas review, debate, and synthesize. Inspired by Karpathy's llm-council.
Language: Go
Author: jtsilverman
18 AI personas deliberate your hardest decisions across multiple LLM providers. Aristotle, Feynman, Kahneman, Torvalds & more — structured multi-round deliberation with genuine model diversity. One command: /council
Language: Shell
Author: 0xNyk
A local web app that automatically generates AI agent harnesses by leveraging a 3-stage LLM Council process — Independent Design, Peer Review, and Chairman Synthesis.
Language: Python
Author: choijunho-AIDeveloper
Compares results of popular LLMs with an AI judge deciding the best response for your query.
Author: Mayank-Madaan-613
Open-source multi-model AI deliberation framework. 5 LLMs debate, peer-review, and synthesize. Better decisions than any single model. Built with MAMA.
Author: OliWoods-Org
LLM Council Plus — Multi-model AI deliberation system with 3-stage council process.
Language: Python
Author: DmitryBMsk
Threaded, follow-up-friendly version of llm-council with conversation memory and full-session context.
Language: Python
Author: FersaceHernandez
Hybrid local/cloud Ollama app that runs your prompt through multiple LLMs, has them critique and rank each other, and returns a single streamed final answer synthesized by a chairman model.
Language: Python
Author: mabuya02
A council of LLMs for hard decisions, inspired by the MAGI of Neon Genesis Evangelion. Deliberate, and synthesise — blind peer review.
Language: Python
Author: jason-chao
An open-source multi-agent terminal workstation mixing local and cloud LLMs for agentic research and system evaluation.
Language: Python
Author: aayushbhaskar
A Claude Code framework for multi-llm planning and development agents.
Language: Python
Author: sherifkozman
Projects targeted at a specific vertical or decision domain. Some entries also appear in other sections where they fit.
AI-powered Design Authority — 5 specialist agents evaluate architecture decisions and deliver structured rulings in under 60 seconds.
Language: Python
Author: JustinNarracott
AI Presidential Briefing: Daily knowledge synthesis system with memory layer, LLM council, and LinkedIn post generation.
Language: Python
Author: Aayushm24
Structured adversarial review for business ideas using a multi-provider LLM council. Inspired by Karpathy's LLM Council and pAI (Poggio Lab, MIT). Critics by design, not by request.
Language: Python
Author: piperod
Personal AI Council — 5 LLM models debate SAP questions. Built with FastAPI, React, Grok API. Free to run locally.
Language: Python
Author: abbhisap
AI Investment Advisor (v1.0): Professional-grade 7-Agent Swarm. Integrates Council Debate mechanism, pgvector semantic memory, and multi-tier LLM routing. A self-optimizing quantitative ecosystem built with Clean Architecture, DSPy, and 75%+ test coverage.
Language: Python
Author: neohsiung
Analyze A-share market trends with AI-driven insights, simulating a top fund's decision-making process through advanced multi-agent collaboration.
Language: CSS
Author: Sunnil07
Boule (Βουλή) — Multi-agent AI council simulating corporate governance. CEO, CFO, CTO and department heads deliberate across iterative cycles with dynamic skills, random events, KPI tracking, and persistent memory. Powered by local LLMs (Ollama/LM Studio) or mock mode. CLI + FastAPI web UI.
Language: Python
Author: aatel-license
Fair Gig Guardian — AI-powered platform that analyzes gig economy contracts to detect potentially unfair clauses. Uses a multi-agent AI council (Auditor, Debate, Judge) to evaluate contract terms and generate fairness scores and actionable insights for gig workers.
Language: TypeScript
Author: Rak2k6
MediCouncil: Multi-Agent LLM Council for Symptom Triage.
Language: Jupyter Notebook
Author: Gayathri05SK
GraphRAG + 4-agent LLM council for biomedical research. Find contradictions, validate hypotheses, and get confidence-scored answers, cited to real papers. Built on Neo4j, LangGraph, Groq & OpenRouter. Also listed under Science & Research.
Language: Python
Author: al1-nasir
NameForge: AI-powered startup name evaluator with LLM Council deliberation. Evaluates names across legal, domain, social, linguistic, strategic & financial dimensions.
Language: TypeScript
Author: fsztpartners
A Gemma-based chatbot app that has multiple bots impersonating famous philosophers from different schools.
Language: Python
Author: dimitreOliveira
A council of 5 ultra-small language models that debate philosophical questions with rebuttals, alliances, achievements, and dramatic visual theater.
Language: Python
Author: zeon01
4 LLMs work together to generate a report based off of scientific papers without hallucinations.
Language: Python
Author: Josephcc2
GraphRAG + 4-agent LLM council for biomedical research. Find contradictions, validate hypotheses, and get confidence-scored answers, cited to real papers. Built on Neo4j, LangGraph, Groq & OpenRouter. Also listed under Medical.
Language: Python
Author: al1-nasir
Can't afford a 7-figure CISO? Assemble a 6-member AI security council instead. Six models, six personas, independent deliberation, multi-dimensional scoring, a Chief Arbiter verdict, and a board-ready PDF. Inspired by Karpathy's LLM Council, built for cybersecurity. Runs locally.
Language: Python
Author: KunalCyber
Offline-first, AI-powered strategic advisory application that simulates a multi-agent council of history's greatest strategic thinkers.
Language: Python
Author: gtm-k
AI-powered supply chain decomposition and risk analysis — council of 5 LLM personas identifies every component, material, and supplier in a product's supply chain.
Language: TypeScript
Author: blackswanworldsim
Browser-based LLM evaluation tool. Test prompts and models on your own data with multi-model judge council, cost tracking, and ranked results.
Language: TypeScript
Author: harshitleads
Meeting Minutes of the AI Council — Ten Leading AI Models Deliberate on the Human-Machine Trust Crisis Triggered by the Claude Mythos. A transparent experiment for human-AI symbiosis governance. (本项目和会议都是在中文环境下完成，若想看到英文版请自行翻译。)
Author: elookto
Multi-LLM consensus extension for Gemini CLI. Inspired by Andrej Karpathy's llm-council.
Language: TypeScript
Author: theerud
Drawing inspiration from Andrej Karpathy's LLM Council, this is an implementation for coding. LLMs evaluate each other and generate the best result rather than modern day IDEs where only one model is chosen.
Language: Python
Author: ibrahimansr
An LLM council that reviews your coding agent's every move.
Language: TypeScript
Author: usetig
PolyCouncil is an open-source multi-model deliberation engine for LM Studio. It runs multiple LLMs in parallel, gathers their answers, scores each response using a shared rubric, and produces a final, consensus-driven result.
Language: Python
Author: TrentPierce
Coven: Offline LLM Council.
Language: Python
Author: pierluigi-failla
A local multi-agent AI debate system where councils of AI personalities deliberate, argue, and reach consensus — running privately on your own hardware via Ollama.
Language: JavaScript
Author: JonahCrut
An LLM council made to thoroughly critique ideas, providing scores, appraisals, & criticisms. Run locally or with LLM providers.
Language: Python
Author: CaptnJayce
Run a council of local LLMs that debate, critique, and synthesize — no API keys needed.
Language: Python
Author: JitseLambrichts
MCP server that gives Claude a review council — other LLMs fact-check responses before you see them.
Language: Python
Author: alterego-987
Multi-LLM group chat MCP server — tool-agnostic, provider-agnostic, token-transparent.
Language: Python
Author: harneet2512
Multi-agent LLM deliberation engine. Create AI agents with custom personalities, assemble councils, and watch them debate, rank, and synthesize answers. Supports OpenAI, Anthropic, Gemini, Groq, Mistral, xAI & any OpenAI-compatible endpoint. Fully client-side for Android and web.
Language: TypeScript
Author: jabezpauls
Projects by the index maintainer that riff on the LLM Council pattern.
Hypothesis-testing variant of the LLM council pattern — 3 diverse models (Sonnet, Grok, MiniMax) via OpenRouter, Sonar grounding, Typst report.
Language: Python
Author: danielrosehill
Method-specific council that evaluates a user-presented decision through seven formal decision-making frameworks — Pros & Cons, Weighted Decision Matrix, 10/10/10, Pre-Mortem / Inversion, SWOT, Six Thinking Hats, and WRAP. One framework per council member; Chairman synthesises across frameworks into a decision dossier with convergent signal, named disagreements, confidence level, and kill criteria. Typst PDF output.
Language: Python
Author: danielrosehill
LLM Council app — grounded multi-model deliberation with peer review and synthesis.
Language: Python
Author: danielrosehill
Domain-specific council for house and apartment hunting. User provides a free-text brief; a five-persona council (Mortgage Expert, Home-Ownership Project Expert, Renovation Specialist, Lifestyle Advocate, Risk Auditor) debates and cross-reviews, then a Chairman synthesises a Search Spec — with an explicit "Acceptable Compromises" section naming what to give up and why. Typst PDF output.
Language: Python
Author: danielrosehill
Template based on Karpathy's llm-council, modified to use a single LLM with six personality-based system prompts (Logical Thinker, Creative Solver, Pessimist, Optimist, Connector, Unconventional) instead of multiple providers. Includes Typst PDF report and Edge TTS podcast digest outputs.
Language: Python
Author: danielrosehill
Voice-first, batch-oriented remix of Karpathy's llm-council. MP3 braindump → STT → cleanup → agent parses a shared Context plus Q1..Qn → each question fans out to a council of models (via OpenRouter) with review + Chairman synthesis → aggregator → Typst PDF report. Ramble in, typeset PDF out.
Language: Python
Author: danielrosehill
LLM Council works together to answer your hardest questions.
Language: Python
Author: karpathy
Claude Code plugin to consult multiple AI coding agents (Gemini, OpenAI, Grok) for diverse perspectives.
Language: Shell
Author: hex
Claude Code Master Prompt: LLM Council for High-Stakes Technical Decisions.
Author: russell0
Multi-LLM council protocol SDK.
Language: Python
Author: peteski22
A decision operating system for high-stakes choices — business, strategy, career. Simulates disagreement, stress-tests assumptions, and converges on what actually holds up. Claude Code skill inspired by Karpathy's autoresearch + LLM council.
Language: Python
Author: harshilmathur
Adversarial AI Council skill for Claude Code — 5 subagents stress-test your decisions. Inspired by Karpathy's LLM Council.
Author: charlomrt-boop
Claude Code skill: Run decisions, code, and plans through a council of 5 AI advisors with anonymous peer review. Based on Karpathy's LLM Council.
Author: ngmeyer
Stress-test decisions with a 16-persona council. A Claude Code skill for PMs, founders, and builders facing wicked problems.
Author: mshadmanrahman
A council of 5 personas to help you take decisions. Choose between 5 free AI models. Based on Ole Lehmann's LLM_Council skill for Claude.
Author: icesixxx-gif
A skill for Claude Code that enables brainstorming with other LLMs (ChatGPT, Gemini) before presenting the implementation plan to the user.
Language: Python
Author: gcpdev
An AI skill that turns any request into a multidisciplinary senior council — profiles, perspectives, consensus, and a clear action. Works with Claude, GPT-4, Gemini, Llama, and any LLM.
Author: efesodavila
A take on Andrej Karpathy's LLM Council, where you are the ultimate head chairman making decisions in the real world. A great tool to prompt multiple models at the same time to compare output quality.
Language: Python
Author: laceyp99
Maintained by Daniel Rosehill