For AI labs and content buyers
Verification is free. Documentation is the product.
Checking a LicenseFoundry credential costs nothing and never calls us — that is deliberate, and it is permanent. What labs pay for is the other half: a durable, independently verifiable record of what they ingested, for the moment someone asks.
The EU AI Act's rules for general-purpose AI require providers to honour rights reservations and publish a summary of their training data — in force now, with enforcement tightening through 2026. Read the compliance guide →
Two paths for labs.
Cryptographic verification of any LicenseFoundry credential runs locally in your SDK against cached JWKS — milliseconds per check, in-process, no metering, no API key, no network call to us. Free for any verifier.
import { verify } from "@licensefoundry/sdk";
const result = await verify(credentialJwt);
if (!result.valid) throw new Error(result.reason);
Python SDK (licensefoundry on PyPI) has the same shape.
Labs have two jobs. The first is always: verify what comes in. The second is conditional on shipping models or generative output: document what goes out.
Path A · AUDIT-only
For labs that ingest credentialed content but don't ship credentialed output.
- Choose your AUDIT tier: Audit Starter $2,199/mo, Audit Growth $6,599/mo, or Audit Enterprise (quoted, $16,599+/mo)
- Unlimited verification (free for everyone)
- Continuous audit trail across your ingested corpus
- Quarterly documentation reports (depth scales with tier)
- No publisher subscription needed
Path B · AUDIT + Publisher bundle
For labs that also ship credentialed output — models, fine-tunes, generative results.
- Subscribe to AUDIT (any tier) for ingestion oversight
- Subscribe to a publisher bundle (Starter / Growth / Scale / Enterprise) for Credentials + Standard Licenses on outputs you ship — ASSESS is free for everyone, no subscription required
- Two subscriptions, two invoices — both predictable monthly
- Each side stands on its own clean tier — no special "bundle SKU"
- Premium and Bespoke LICENSE attestations remain à-la-carte
Why this is cleaner than a combined bundle: each side has different usage patterns. Lab ingest grows with training data; output issuance grows with commercial product shipped. Two subscriptions scale independently. One invoice covering both forces compromises in either ingestion or output capacity.
More than compliance. A commercial advantage.
The defensive case is straightforward — when regulators or counsel ask, you have signed evidence. The commercial case is what most prospects miss: credentials change what each side can sell and what they can charge for. The same artefact that protects you when something goes wrong also widens the deal you can sign when things go right.
For content platforms
Charge a premium for credentialed catalogues.
A licensed dataset that ships with a signed credential trail is worth materially more to an AI lab than the same dataset on a "trust us" basis. Labs pay more for training data they can defend — and you're the one selling them that defensibility. The same credentials also win bilateral deals against competitors who can only offer their own assertion. Same catalogue, higher take rate.
For AI labs
Sell certainty to your enterprise customers.
When your enterprise customers' procurement teams ask "what was this model trained on, and was it cleared?", today your answer is "trust us." With a credentialed training corpus, your answer is a signed report from an independent third party — verifiable by their legal team without trusting either of us. That converts the conversation from defending an objection to justifying a premium tier.
Compliance protects the floor of the business. Credentialed inputs raise the ceiling — for issuers, by widening what their catalogue is worth; for labs, by letting them charge enterprise customers a premium for provable provenance.
Talk to us about audit.
Verification stays free whether or not you ever speak to us. If you need documented oversight of your training-corpus rights, that is the conversation to have.