Technical documentation.
Implementation guides for each layer of the Lemma architecture.
-
Define Your Domain as a Schema
Model how your AI retrieves and clusters knowledge — bucket ages, risk scores, regions — with typed schemas and normalization. Register ZK circuits and generators so every fact traces back to its source.
Read → -
Encrypt Everything, Expose Nothing
How @lemmaoracle/sdk keeps every document AES-256-GCM encrypted on the client so your AI never touches raw PII — only docHash and CID are exposed as stable anchors for provenance.
Read → -
Prove Facts with Zero Knowledge
Turn business rules like 'over 18' or 'revenue above threshold' into machine-checkable facts. Each proof is permanently recorded with its circuit and generator.
Read → -
Disclose Only What AI Needs
Selective disclosure lets holders reveal just the attributes your model requires, while the link to the original issuer signature stays intact.
Read → -
Query Verified Attributes
Ask 'users over 18 in Japan' and get back attributes with full provenance — proof status, schema, issuer, generator, and verification method — ready for your RAG policy layer.
Read → -
Provenance That Never Disappears
Document commitments, schemas, issuers, and ZK verification results are anchored on-chain. Your RAG index can be rebuilt, your embeddings re-computed — the provenance layer stays permanent.
Read →
Data stays in place.
Only proofs travel.
The source mints a proof, only the necessary parts are selectively disclosed, and a verifier checks them with a public key. Through all three phases, raw data never moves.
Mint the proof at the source
Embed the Lemma SDK at the source. Issue attribute proofs for data, models, and authority. Raw data never leaves.
Reveal just enough
Selectively disclose only the facts a verifier needs — authority, attribute, integrity. Raw data is never touched.
Anyone can confirm
Anyone, anytime, verifies authenticity with a public key. Optionally anchor as tamper-evident audit attestations.
Proprietary cryptography,
three cores.
Three core technologies make the trust infrastructure work — establishing “it’s the same thing,” “disclose only this,” and “no one can rewrite it,” all without touching the original data.
Proof
Turn business rules like “18 or older” or “revenue over threshold” into machine-verifiable cryptographic facts. Prove a statement’s truth without touching the original.
Disclosure
Keeping the binding to the original issuer signature, the holder reveals only the attributes a model needs. Minimize what AI receives as input.
Provenance
Anchor the document commitment, schema, issuer, and ZK verification result on-chain. Fix it as provenance that can be neither rewritten nor erased.