Compute Receipts

Understanding cryptographic proofs of computation

What is a Compute Receipt?

A compute receipt is a cryptographic proof that attests to the correct execution of a computation. When a validator executes a compute job, they produce a receipt that includes:

  • A commitment to the input data
  • The output of the computation
  • A proof that the computation was executed correctly
  • Metadata about the execution environment

Receipt Structure

Each compute receipt follows a standardized format that enables independent verification:

receipt.json
{
  "version": "0.1.0",
  "job_id": "0x1a2b3c...",
  "input_commitment": "0xabc123...",
  "output_hash": "0xdef456...",
  "proof": {
    "type": "snark",
    "data": "0x..."
  },
  "validator": "0x789xyz...",
  "timestamp": 1699999999,
  "signature": "0x..."
}

Verification Process

Any network participant can verify a compute receipt without re-executing the computation. The verification process involves:

  1. Checking the proof against the stated inputs and outputs
  2. Verifying the validator signature
  3. Confirming the receipt is properly anchored in the consensus layer

Testnet Note

The current testnet uses simplified proof mechanisms for faster iteration. Production-grade proof systems are under development.

Receipt Types

ComputeNet supports different receipt types depending on the verification requirements:

  • Full receipts — Complete cryptographic proofs for maximum security
  • Attestation receipts — Validator signatures for trusted scenarios
  • Aggregate receipts — Batched proofs for efficiency

Use Cases

Compute receipts enable a variety of applications:

  • Verifiable machine learning inference
  • Provable data transformations
  • Auditable batch processing
  • Cross-chain compute verification
NextOverview