OriginTrail is an ecosystem building collective, trusted memory for AI. The core ecosystem technology is the Decentralized Knowledge Graph (DKG), a decentralized, permissionless network of nodes, through which both humans and machines can share knowledge, reason together, and preserve context across time.
Modern AI is powerful but ungrounded. It predicts without knowing, hallucinates, forgets what it said, and relies on data controlled by a few centralized platforms. LLMs in particular have an "explainability" problem - why did an LLM respond a certain way, based on what knowledge, coming from which source?
The Decentralized Knowledge Graph (DKG) hosts Knowledge Assets that encode facts, data provenance, and meaning in a tamper-proof way. The network is hosted by a set of independent DKG Nodes. Anyone can run a DKG network Node - organizations and individuals - contributing to the DKG and at the same time building upon their knowledge in a privacy-preserving way. Thus the DKG ensures that no single entity can rewrite, censor, or monopolize the collective memory. Decentralization keeps AI accountable, bias-resistant, and aligned with human diversity.
OriginTrail’s neuro-symbolic approach combines the structure and reasoning of symbolic AI with the creativity and pattern recognition of neural AI. This enables AI systems to think with context, grounding their outputs in verifiable knowledge rather than probabilistic guesses.
The DKG grows through human participation. Researchers, developers, and citizens can all publish, link, and improve knowledge — ensuring that the world’s intelligence is shaped by the many, not the few.
Core operations
- Publishing knowledge: Turning data into structured, verifiable Knowledge Assets
- Knowledge Discovery: Querying, traversing, and monetizing knowledge in the decentralized graph and its paranets
- Trusted sharing: Cryptographically verify authenticity and provenance of knowledge
- Neuro-symbolic Reasoning: Infer new facts based on rules, leveraging graph-based reasoning in combination with LLMs and GenAI models
We encourage developers to try out the DKG Node and build their first DKG based agent with it, to get a feel of what the technology can do.
| Cover image | |||
|---|---|---|---|
| Build your DKG Agent | Begin your journey on the Decentralized Knowledge Graph by staking TRAC and setting up your first DKG Node. This is your entry point into the verifiable knowledge economy | Delegated Staking.png | |
| Contrubute to the DKG | Activate your DKG Node and join the decentralized knowledge economy by staking TRAC. | DKG key concepts.png | dkg-key-concepts.md |
| Learn more about the DKG | Explore the core concepts behind the DKG and how it powers verifiable, intelligent AI. | DKG under the hood.png |
{% hint style="success" %} This site is a constantly updated, work-in-progress official OriginTrail documentation built by the OriginTrail community and core developers.
Find the "Edit on GitHub" button in the upper-right corner and submit your updates as pull requests (or do it directly in the docs GitHub repo).
We appreciate any feedback, improvement ideas, and comments. {% endhint %}