What is codatta?
Codatta is an on-chain knowledge layer that assetifies human- and AI-produced data into ownable, tradable, revenue-sharing assets. It provides a data system (fingerprints, lineage, access control, privacy) and a royalty engine (licensing + automatic payouts), plus a network of human & AI agents (identity, reputation) to source, label, validate, and evaluate data and AI systems.Why it exists
- AI progress is throttled by data cost, quality, and attribution.
- Contributors rarely retain ownership or share in downstream value.
- Buyers want train-now, pay-later / pay-on-results options with trustable provenance.
The Royalty Economy (core)
- Ownership fractions: Contributions + verification + staking establish fractional ownership over data assets.
-
Usage recording & metering:
- Training / fine-tune / evaluation consumption is recorded for lineage and attribution.
- Inference of downstream models derived from licensed data is metered (e.g., requests/tokens/API calls) for settlement.
- Automatic payouts: Smart contracts split royalties to contributors, validators, backers, and the protocol treasury—continuously, per agreed terms.
- Modes supported: pay-as-you-train, TNPL (Train-Now, Pay-Later), and performance-linked pay-on-results (KPIs, SLAs).
In short: better data → better models → more usage → more royalties—and the right people get paid automatically.
What you get
- Ownable data assets (samples + labels + validations) with immutable lineage.
- Royalty-based payouts to contributors and backers whenever assets are used/licensed or when derived models are used in production.
- Pluggable backend to add ownership + royalty sharing to existing platforms without re-architecting their frontends.
Not a marketplace. codatta plugs into platforms and marketplaces that already exist—acting as their knowledge & royalty backend (provenance, policy, metering, and payouts).