About Trace · A Data Foundation initiative
The record of how this generation's models learned what they know.
Trace is the public audit layer for the data behind AI. Every contribution becomes an immutable receipt on a transparent ledger that anyone can check.
For labs
Licensing data without a leap of faith.
Every year, frontier labs license north of half a billion dollars of training data, and most of it ships with a vendor's word as the only audit. Trace is where that word becomes a record.
Every consent acceptance, every TOS version, every metadata update, all written to a public ledger that anyone can verify and no one can quietly rewrite. Sign with confidence. Defend with evidence.
For contributing apps
Trust you don't have to build.
Building a credible audit surface in-house is a serious engineering lift and a permanent target for accusations of bias. Trace is the impartial alternative: a public ledger your customers, regulators, and contributors all check on the same terms.
Send a webhook. Get a branded trust portal at your subdomain. Point procurement teams at it. Your data, your branding, third-party verifiability.
How it stays honest
Three commitments.
Open infrastructure earns trust by what it refuses to do as much as by what it does. Here's the shape of the refusal.
- 01
A transparent ledger
Every receipt is written to a shared, public ledger that stays permanent and tamper-evident by design. No quiet edits, no retconning, no off-the-record backchannels.
- 02
Open by design
No login. No gating. Anyone can verify any
data_id, look up any contributor status, and audit any dataset against the same public source. - 03
Not a content store
Trace holds the receipt, not the asset. Sensitive metadata values stay with the contributing app, and Trace records existence flags only.
Building a contributing app?
Integrate the webhook, get a branded whitelabel portal at your subdomain, and let your customers verify against the same public ledger every other contributing app uses.