Skip to main content
Docs/Core concepts

Execution receipts

Receipts create a verifiable trail for skill runs, payments, providers, and output hashes.

Last updated July 2, 2026

Receipt data

Every execution should leave behind enough data to prove what ran without exposing private prompt content or sensitive user input. The current app stores receipts in the database; the production protocol design writes compact receipt proofs to Solana accounts.

AttributePurpose
Skill IDIdentifies which skill was run.
Executor walletConnects the run to the paying public wallet.
Fee paidSupports accounting and revenue distribution.
ProviderShows which LLM route produced the output.
Output hashVerifies the returned result without storing full private output on-chain.
TimestampSupports audit, analytics, and dispute review.

Privacy model

Keep receipts compact

Never write private prompts, user secrets, API keys, documents, or full LLM outputs to a public ledger. Store hashes and references only.

NextArchitecture