Skip to main content

Introduction to Aragora

Aragora is the control plane for multi-agent vetted decisionmaking across organizational knowledge and channels. It orchestrates 30+ agent types to debate your organization's knowledge (documents, databases, APIs) and deliver defensible decisions to any channel (Slack, Teams, Discord, voice).

Why Multi-Agent Debate?​

Single AI models have inherent biases and blind spots. Multi-agent debate addresses this by:

  • Diverse perspectives: Different models (Claude, GPT, Gemini, etc.) bring unique reasoning approaches
  • Self-correction: Agents critique and improve each other's responses
  • Consensus building: Final answers represent agreement across multiple AI systems
  • Transparency: The vetted decisionmaking process is fully auditable

Core Features​

Heterogeneous Agents​

Aragora supports agents from multiple providers:

ProviderModelsStrengths
AnthropicClaude 3.5, Claude 4Reasoning, safety
OpenAIGPT-4, GPT-4oGeneral knowledge
GoogleGemini Pro, Gemini FlashMultimodal, speed
xAIGrokReal-time knowledge
Open ModelsLlama, Mistral, DeepSeekCost efficiency

Structured Debates​

Debates follow configurable protocols:

const protocol = {
  rounds: 3,
  phases: ['opening', 'critique', 'revision', 'vote'],
  consensus: {
    type: 'supermajority',
    threshold: 0.75,
  },
};

Memory & Learning​

Aragora maintains multi-tier memory for continuous improvement:

  • Fast memory: Immediate context (session-scoped)
  • Medium memory: Cross-session patterns
  • Slow memory: Long-term strategic knowledge
  • Glacial memory: Archived historical insights

Knowledge Mound​

The Knowledge Mound is Aragora's intelligent knowledge base that:

  • Stores debate outcomes and evidence
  • Supports semantic search across findings
  • Enables knowledge sharing between workspaces
  • Federates across organizational boundaries

Use Cases​

Security Analysis​

Deploy multiple AI agents to analyze code, infrastructure, or policies for vulnerabilities. Each agent brings different security expertise.

Research Synthesis​

Have agents debate and synthesize findings from multiple sources, ensuring comprehensive coverage and accurate conclusions.

Decision Support​

Complex business decisions benefit from multi-perspective analysis and structured vetted decisionmaking.

Content Review​

Quality assurance for generated content, documentation, and technical writing.

Getting Started​

Ready to start? Follow these guides:

  1. Quickstart - Run your first debate in minutes
  2. Installation - Set up Aragora in your environment
  3. First Debate - Step-by-step tutorial

Architecture Overview​

                    ┌─────────────────────────────┐
                    │       Aragora Server        │
                    │  ┌─────────────────────────┐│
   Clients ────────▶│  │    Debate Orchestrator  ││
                    │  └──────────┬──────────────┘│
                    │             │               │
                    │  ┌──────────▼──────────────┐│
                    │  │      Agent Pool         ││
                    │  │ ┌─────┐┌─────┐┌─────┐   ││
                    │  │ │Claude││ GPT ││Gemini│  ││
                    │  │ └─────┘└─────┘└─────┘   ││
                    │  └──────────┬──────────────┘│
                    │             │               │
                    │  ┌──────────▼──────────────┐│
                    │  │   Memory & Knowledge    ││
                    │  │   ┌────────────────┐    ││
                    │  │   │ Knowledge Mound│    ││
                    │  │   └────────────────┘    ││
                    │  └─────────────────────────┘│
                    └─────────────────────────────┘

Community​

License​

Aragora is available under AGPL-3.0 for open-source use, with commercial licenses available for enterprise deployments.