Lemma — Verifiable Provenance For Ai Agents
@lemmaoracle
AI agents make decisions you can't audit. Lemma fixes that with zero-knowledge proofs over confidential documents — selective disclosure, on-chain provenance, no plaintext exposure. Query cryptographically verified attributes from agents in production. Built for regulated industr
Overview
What is Lemma — Verifiable Provenance For Ai Agents?
Lemma is a cryptographically verified truth layer for AI agents. It enables reasoning over confidential data via zero-knowledge proofs, selective disclosure, and tamper-evident provenance — while raw content stays encrypted. Every attribute an AI reads through Lemma carries permanent provenance: who issued it, which schema defined it, how it was proven, and where its verification record is anchored.
How to use Lemma — Verifiable Provenance For Ai Agents?
Install the MCP server by adding a configuration entry to Claude Desktop or any MCP-compatible agent. Use the command npx -y @lemmaoracle/mcp and set the LEMMA_API_KEY environment variable to your API key.
Key features of Lemma — Verifiable Provenance For Ai Agents
- Cryptographically verified provenance for every attribute.
- Zero-knowledge proofs for confidential data reasoning.
- Selective disclosure of attributes while raw content stays encrypted.
- Tamper-evident record anchored to a verification ledger.
- Issuer, schema, proof method, and verification location permanently attached.
Use cases of Lemma — Verifiable Provenance For Ai Agents
- AI agent queries verified facts from a confidential dataset without exposing raw data.
- Agent selectively discloses only specific attributes from a verified credential.
- Agent or application verifies the issuer and schema of
