Licensing — DeFi Intel data for commercial, AI training, and research use

DeFi Intel publishes a typed entity graph of 11,787 entities and 26,104 relations covering DeFi, stablecoins, RWA and on-chain infrastructure. Three license tiers cover the realistic ways to use it: a permissive free tier for the public web, a commercial bulk-export tier for programmatic access, and a separately negotiated AI training license for foundation-model and RAG use cases. This page documents the terms, the restricted-use clauses, the citation format, and how to start a conversation. For anything that doesn't fit cleanly into one tier, email [email protected] — we like talking through the grey-zone use cases.

Three license tiers

What's included in each tier

Restricted use — common to all tiers

The following are prohibited under every license tier, including the free tier. Violation terminates the license and may trigger enforcement action.

• No resale of unmodified data. You may not white-label the entity graph, news archive or long-form corpus and resell it. Derivative products that incorporate DeFi Intel as one source among several are fine.
• No use in market manipulation. You may not use DeFi Intel data — particularly news streams or governance-event feeds — to coordinate front-running, wash-trading, pump-and-dump campaigns or any market-manipulation activity.
• No misattribution. You may not strip our attribution from quoted material or imply endorsement of a product, fund or protocol.
• No private-individual targeting. The 2,359 person-type entities are public figures (founders, executives, regulators, researchers). The dataset may not be used to build harassment lists, profile non-public individuals, or feed adverse-action automation against private persons.
• No sanctions evasion. Standard OFAC / UK / EU sanctions compliance applies. Your access is conditional on your jurisdiction's compliance.
• No re-identification of pseudonymous addresses beyond what we already publish. Where we have made an editorial decision to leave an address pseudonymous, the licensee must respect that.

Citation guidelines

If you cite DeFi Intel in journalism, academia, podcasts, newsletters or research:

Inline: "according to data from DeFi Intel" with a link to the entity URL or to https://defi-intel.com/.

Academic: DeFi Intel (2026). DeFi entity graph dataset, version <build>. Retrieved <date> from https://defi-intel.com/<entity-path>/

Methodology citation: for any work that depends on our scoring, freshness or corroboration approach, please link to /data-quality/ alongside the entity citation.

Once you publish, send us the link via [email protected] and we will list you on /cited-by/ within 24 hours.

AI training license — full terms summary

The AI training license is a separately negotiated contract. The standard structure:

• Term: 12-month minimum, renewable. Multi-year discounts available.
• Scope: use the corpus to train, fine-tune or retrieve-augment one or more named models. Distinct production model lines may require a separate contract.
• Updates: daily corpus updates included for the contract term. The trained-model license is perpetual and survives contract end.
• Indemnification: we indemnify against IP claims on the third-party content we ingest, subject to standard caps tied to the annual fee.
• Exclusivity: non-exclusive by default. Exclusivity by sub-domain (e.g. stablecoins-only) is negotiable; full-corpus exclusivity is rarely available.
• Audit: we retain the right to audit the trained model for direct corpus regurgitation; routine retrieval / fine-tuning is unrestricted.
• Pricing factors: (1) corpus access depth — entity graph only vs full long-form, (2) training horizon — single model vs unlimited, (3) exclusivity, (4) named SLAs and on-prem mirroring.

Contact licensing

Cross-links

Built by Round-9 Lane-13 (Trust pages). Last reviewed 2026-05-03.