Documentation
Provenance is the trust layer for generative AI. Our APIs trace model behavior, verify media authenticity, and authorize agent actions.
The sandbox ships with pre-provisioned keys and seeded fixtures, so you can make your first call without an account. Start with the Quickstart.
Start here
Browse by product
Trace
Overview · Traced inference · Attribution traces · Canary policies · Kill switch · Neuron registry · Self-hosted sidecar · Neuron-CI diffs
Attest
Attestation bundles · Video attestation · Verification · Risk scoring · Synthetic-media screening · Insurance claims · Field inspections
Authorize
Agent authorization · Policies · Verifying tokens · Delegation mandates · Decision ledger · Agent payments · MCP server
Platform
Authentication · Errors · Webhooks · Embeds · Transparency log · Billing
Tools
- SDKs and verifiers: TypeScript and Python clients, plus verifier-only libraries for Python and Go.
- MCP server: connect Claude or any MCP client to Provenance.
- API reference: every route, generated from the OpenAPI spec.