llm-council

The LLM Council

$ council run architect "Design a mass hallucination prevention system"

                    ╔══════════════════════════════════════════════════════════╗
                    ║             ⚖️  THE LLM COUNCIL CONVENES  ⚖️              ║
                    ╚══════════════════════════════════════════════════════════╝

      ┌─────────────────┐      ┌─────────────────┐      ┌─────────────────┐
      │  ┌───────────┐  │      │  ┌───────────┐  │      │  ┌───────────┐  │
      │  │ ╭───────╮ │  │      │  │ ╭───────╮ │  │      │  │ ╭───────╮ │  │
      │  │ │GPT5.2│ │  │      │  │ │ CLAUDE│ │  │      │  │ │GEMINI │ │  │
      │  │ ╰───────╯ │  │      │  │ ╰───────╯ │  │      │  │ ╰───────╯ │  │
      │  │   ◉ ◉     │  │      │  │   ◉ ◉     │  │      │  │   ◉ ◉     │  │
      │  │    ⌣      │  │      │  │    ▽      │  │      │  │    ○      │  │
      │  └───────────┘  │      │  └───────────┘  │      │  └───────────┘  │
      │    JUDGE #1     │      │    JUDGE #2     │      │    JUDGE #3     │
      └────────┬────────┘      └────────┬────────┘      └────────┬────────┘
               │                        │                        │
               │ "I propose we use      │ "Actually, I must      │ "Interesting, but
               │  a vector database..." │  respectfully disagree" │  what about...?"
               │                        │                        │
               └────────────────────────┼────────────────────────┘
                                        ▼
                         ┌──────────────────────────────┐
                         │     🔥 ADVERSARIAL DEBATE 🔥   │
                         │                              │
                         │  GPT5.2: "Your approach has  │
                         │          a cold start issue" │
                         │                              │
                         │  CLAUDE: "Fair, but yours    │
                         │          doesn't scale"      │
                         │                              │
                         │  GEMINI: "Both valid. What   │
                         │          if we combine..."   │
                         └──────────────┬───────────────┘
                                        ▼
                         ┌──────────────────────────────┐
                         │      ✅ VERDICT REACHED ✅     │
                         │                              │
                         │   Synthesized best ideas     │
                         │   Schema-validated output    │
                         │   Confidence: 94%            │
                         └──────────────────────────────┘

[Council] Task completed in 45.2s | 3 judges | 2 debate rounds | Cost: $0.12

A Multi-LLM Council Framework that orchestrates multiple LLM backends to enable adversarial debate, cross-validation, and structured decision-making.

Why Use a Council?

Single-model outputs have blind spots. By running multiple models in parallel and having them critique each other, the council:

• Catches errors that any single model might miss
• Reduces hallucination through cross-validation
• Produces higher-quality outputs via adversarial refinement
• Validates structure with JSON schema enforcement and retry logic

Features

Feature	Description
Multi-Model Council	Run Claude, GPT-4, and Gemini in parallel via single OpenRouter key
Adversarial Critique	Built-in critique phase identifies weaknesses and blind spots
Schema Validation	JSON schema validation with automatic retry for structured outputs
Provider Agnostic	Swap between OpenRouter, direct APIs, or CLI-based providers
Health Checks	Preflight provider health checks with latency tracking
Graceful Degradation	Automatic retry, fallback, and skip strategies for failures
Artifact Store	Persistent storage of drafts with tiered summarization
Secret-Safe Logging	Redaction pipeline prevents credential leakage

Requirements