Cortensor Docs¶
Last updated: 2026-06-28
Cortensor is a decentralized AI execution and verification network. The stack combines on-chain coordination, router nodes, miner execution, oracle/validator workflows, and product surfaces such as the Portal, API gateway, dashboard, Corgent, Bardiel, and PyClaw-facing integrations.
Cortensor routes inference across heterogeneous hardware, supports open-weight models, validates outputs through Proof of Inference and Proof of Useful Work concepts, and exposes both Web2-style APIs and Web3/agentic interfaces. The stack is organized around protocol contracts, node execution, router APIs, product surfaces, ecosystem programs, and operational references.
Current Model¶
Protocol plane
Contracts coordinate roles, node pools, sessions, task queues, payments, privacy, stats, and validation.
Execution plane
Routers, miners, oracles, validators, and support tools run inference and verification workflows.
Product plane
Portal, API gateway, mining dashboard, Corgent, Bardiel, and agent apps make the network usable.
Start Here¶
- Start with What Is Cortensor Today for the stack-level picture.
- Use Architecture At A Glance for the protocol and execution flow.
- Use Products for Router Node, Portal, Corgent, Bardiel, PyClaw, agent apps, and dashboard surfaces.
- Use Getting Started for developer, router, miner, and Web3 paths.
- Use Reference for terminology and the model catalog.
Core Pillars¶
| Pillar | How Cortensor supports it |
|---|---|
| Decentralized AI inference | Router nodes accept work, miners execute it, and sessions/queues coordinate task lifecycle. |
| Universal AI accessibility | Hosted Portal/API access and self-hosted router access give developers Web2 and Web3 paths. |
| Community-powered compute | Node roles, model capacity, dashboard visibility, and incentive surfaces describe how operators participate. |
| Verifiable AI | PoI, PoUW, validation routes, Corgent, and Bardiel show how results can be checked before action. |
| Open model ecosystem | The model catalog separates model IDs, engines, capacity, and deployment availability. |
Site Structure¶
| Section | What it covers |
|---|---|
| Overview | Narrative, architecture, and current stack context. |
| Getting Started | Reader paths, node setup, router setup, Web2 API, and Web3 SDK reference. |
| Core Concepts | Roles, sessions, PoI/PoUW, private inference, and multi-layer network concepts. |
| Stack | Contracts, runtime, router flows, infra, and ops. |
| Products | Router Node, Portal, Dashboard, Corgent, Bardiel, PyClaw, and agent apps. |
| Community & Ecosystem | Tokenomics, use cases, hackathons, contributors, and partnerships. |
| Roadmap | Historical roadmap, active 2026 plan, status matrix, and long-term vision. |
| Reference | Glossary and model catalog. |