Home
Dissensus is a multi-agent AI dialectical debate system where three specialized AI agents engage in structured, truth-seeking argumentation on any topic. Unlike single-model AI responses, Dissensus leverages adversarial debate to surface hidden assumptions, challenge weak arguments, and forge consensus from chaos.
The system combines the rigor of formal debate with the power of large language models, creating a unique platform for exploring complex topics through structured disagreement.
| Environment | URL | Description |
|---|---|---|
| Landing Page | https://dissensus.fun | Project overview and introduction |
| Debate App | https://app.dissensus.fun | Live debate interface — start debating now |
| Agent | Role | Color | Function |
|---|---|---|---|
| CIPHER 🔴 | The Skeptic | Red | Red-team auditor who finds weaknesses, risks, and flaws |
| NOVA 🟢 | The Advocate | Green | Visionary optimist who finds opportunities and catalysts |
| PRISM 🔵 | The Synthesizer | Blue | Neutral referee who weighs both sides and delivers the verdict |
Together, they form a complete dialectical system: thesis (NOVA), antithesis (CIPHER), and synthesis (PRISM).
| Feature | Implementation |
|---|---|
| Backend | Node.js + Express |
| Streaming | Server-Sent Events (SSE) for real-time debate chunks |
| AI Providers | Multi-provider support: DeepSeek, Google Gemini, OpenAI |
| Frontend | Vanilla JavaScript with EventSource API |
| Deployment | VPS-ready with nginx reverse proxy support |
| Security | Helmet headers, rate limiting, XSS prevention |
- Getting-Started — Installation, configuration, and first debate
- Architecture — Technical overview and data flow
- The-Three-Agents — Deep dive into CIPHER, NOVA, and PRISM
- Debate-Methodology — The 4-phase dialectical process
"Truth emerges from the clash of opposing viewpoints, not from echo chambers."
Dissensus is built on the belief that the best way to understand complex topics is through structured adversarial debate. By forcing AI agents to defend opposing positions and challenge each other's reasoning, we surface insights that single-perspective analysis would miss.
The system is designed for:
- Researchers exploring controversial or complex topics
- Investors evaluating opportunities from multiple angles
- Decision-makers who need to understand both sides before committing
- The curious who want to see how AI agents reason through disagreement
This wiki documents the Dissensus Engine — the core debate system powering the platform. The engine is open-source and designed for easy deployment on any Node.js-capable hosting platform.
Built with ⚡ by the Dissensus team
Where Disagreement Forges Truth