Generate Proof

Request

proofType

string

required

Type of proof to generate. Options:

performance - Prove agent performance metrics
capability - Prove agent capabilities
transaction - Prove transaction validity
compliance - Prove regulatory compliance
data_integrity - Prove data hasn’t been tampered with
execution - Prove code execution occurred correctly

circuitType

string

default:"zk-SNARK"

Zero-Knowledge circuit type. Options: zk-SNARK, zk-STARK

publicInputs

object

required

Public inputs that will be revealed in the proof

{
  "timestamp": 1706380800,
  "version": "1.0",
  "merkleRoot": "0x7d8f9e8a..."
}

privateInputs

object

required

Private inputs used in computation but not revealed

{
  "secretData": "encrypted_blob",
  "privateKey": "0x...",
  "sensitiveMetrics": {...}
}

assertions

array

required

List of statements to prove

[
  "accuracy > 0.95",
  "latency < 100ms",
  "data_not_modified",
  "authorized_access_only"
]

metadata

object

Additional metadata for the proof

{
  "agentId": "agent_123",
  "taskId": "task_456",
  "description": "Model accuracy proof"
}

curl -X POST https://api.treeship.dev/v1/proofs/generate \
  -H "Authorization: Bearer YOUR_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "proofType": "performance",
    "circuitType": "zk-SNARK",
    "publicInputs": {
      "timestamp": 1706380800,
      "modelVersion": "v2.1",
      "datasetHash": "0x7d8f9e8a7b6c5d4e"
    },
    "privateInputs": {
      "testResults": [0.97, 0.96, 0.98, 0.95],
      "trainingData": "encrypted_blob",
      "hyperparameters": {
        "learning_rate": 0.001,
        "batch_size": 32
      }
    },
    "assertions": [
      "average_accuracy > 0.95",
      "min_accuracy > 0.90",
      "training_data_private"
    ],
    "metadata": {
      "agentId": "agent_ml_001",
      "description": "ML model performance proof"
    }
  }'

Response

proofId

string

Unique identifier for the generated proof

proofType

string

Type of proof that was generated

circuitType

string

Zero-Knowledge circuit type used

proof

object

The generated Zero-Knowledge proof

Show proof properties

data

string

Base64 encoded proof data

publicInputs

object

Public inputs used in the proof

verificationKey

string

Public key for verifying this proof

circuitHash

string

Hash of the circuit used to generate the proof

assertions

object

Results of each assertion

{
  "average_accuracy > 0.95": true,
  "min_accuracy > 0.90": true,
  "training_data_private": true
}

status

string

Proof generation status: success, failed, partial

generationTime

number

Time taken to generate the proof in milliseconds

proofSize

number

Size of the proof in bytes

metadata

object

Metadata provided in the request

createdAt

string

ISO 8601 timestamp of proof generation

expiresAt

string

ISO 8601 timestamp when proof expires (if applicable)

{
  "proofId": "proof_7d8f9e8a7b6c5d4e",
  "proofType": "performance",
  "circuitType": "zk-SNARK",
  "proof": {
    "data": "0x8f9d3a4e5b6c7d8e9f0a1b2c3d4e5f6g7h8i9j0k1l2m3n4o5p6q7r8s9t0u1v...",
    "publicInputs": {
      "timestamp": 1706380800,
      "modelVersion": "v2.1",
      "datasetHash": "0x7d8f9e8a7b6c5d4e"
    },
    "verificationKey": "0x2a3b4c5d6e7f8g9h0i1j2k3l4m5n6o7p8q9r0s1t2u3v4w5x6y7z8a9b0c1d2e",
    "circuitHash": "0x9e8d7c6b5a4f3e2d1c0b9a8f7e6d5c4b"
  },
  "assertions": {
    "average_accuracy > 0.95": true,
    "min_accuracy > 0.90": true,
    "training_data_private": true
  },
  "status": "success",
  "generationTime": 156,
  "proofSize": 192,
  "metadata": {
    "agentId": "agent_ml_001",
    "description": "ML model performance proof"
  },
  "createdAt": "2024-01-27T12:30:00Z",
  "expiresAt": "2024-02-27T12:30:00Z"
}

Proof Types

Performance Proof

Prove AI model performance without revealing training data:

const performanceProof = await treeship.proofs.generate({
  proofType: 'performance',
  publicInputs: {
    modelId: 'model_123',
    benchmarkDataset: 'imagenet_test'
  },
  privateInputs: {
    accuracyScores: [0.97, 0.96, 0.98],
    trainingData: encryptedData,
    modelWeights: privateWeights
  },
  assertions: [
    'top1_accuracy > 0.95',
    'inference_time < 50ms',
    'no_data_leakage'
  ]
});

Transaction Proof

Prove transaction validity without revealing amounts:

const transactionProof = await treeship.proofs.generate({
  proofType: 'transaction',
  publicInputs: {
    transactionId: 'tx_789',
    timestamp: Date.now()
  },
  privateInputs: {
    amount: 50000,
    balance: 75000,
    fees: 250
  },
  assertions: [
    'amount > 0',
    'balance >= amount + fees',
    'authorized_sender'
  ]
});

Compliance Proof

Prove regulatory compliance:

const complianceProof = await treeship.proofs.generate({
  proofType: 'compliance',
  publicInputs: {
    regulation: 'GDPR',
    jurisdiction: 'EU'
  },
  privateInputs: {
    userConsents: consentDatabase,
    dataProcessingLogs: auditLogs,
    retentionPeriods: retentionData
  },
  assertions: [
    'explicit_consent_obtained',
    'data_minimization_applied',
    'retention_limits_respected'
  ]
});

Advanced Features

Batch Proof Generation

Generate multiple proofs efficiently:

const batchProofs = await treeship.proofs.generateBatch([
  {
    proofType: 'performance',
    publicInputs: {...},
    privateInputs: {...},
    assertions: [...]
  },
  {
    proofType: 'compliance',
    publicInputs: {...},
    privateInputs: {...},
    assertions: [...]
  }
]);

Proof Composition

Combine multiple proofs into one:

const composedProof = await treeship.proofs.compose({
  proofs: [performanceProof.proofId, complianceProof.proofId],
  composition: 'AND', // All proofs must be valid
  metadata: {
    description: 'Combined performance and compliance proof'
  }
});

Custom Circuits

Use custom Zero-Knowledge circuits:

const customProof = await treeship.proofs.generate({
  proofType: 'custom',
  circuitDefinition: {
    language: 'circom',
    source: circuitSource,
    constraints: 1000000
  },
  publicInputs: {...},
  privateInputs: {...}
});

Use Cases

AI Model Verification

Prove model performance without revealing training data:

// Prove a medical AI model meets accuracy requirements
const medicalProof = await treeship.proofs.generate({
  proofType: 'performance',
  publicInputs: {
    modelType: 'diagnostic_classifier',
    certificationStandard: 'FDA_510k'
  },
  privateInputs: {
    patientData: encryptedPatientTests,
    diagnosisResults: confidentialResults
  },
  assertions: [
    'sensitivity > 0.95',
    'specificity > 0.90',
    'no_bias_detected'
  ]
});

Financial Compliance

Prove financial compliance without exposing transaction details:

const financialProof = await treeship.proofs.generate({
  proofType: 'compliance',
  publicInputs: {
    regulation: 'SOX',
    auditPeriod: '2024-Q1'
  },
  privateInputs: {
    transactions: quarterlyTransactions,
    controls: internalControls
  },
  assertions: [
    'all_transactions_authorized',
    'segregation_of_duties',
    'audit_trail_complete'
  ]
});

Supply Chain Verification

Prove supply chain integrity:

const supplyChainProof = await treeship.proofs.generate({
  proofType: 'data_integrity',
  publicInputs: {
    productId: 'prod_789',
    destination: 'customer_123'
  },
  privateInputs: {
    supplierData: confidentialSupplierInfo,
    logisticsData: shippingDetails,
    qualityData: testResults
  },
  assertions: [
    'authentic_origin',
    'cold_chain_maintained',
    'quality_standards_met'
  ]
});

Best Practices

Never include sensitive data in publicInputs. Use privateInputs for confidential information.

Input Validation: Always validate inputs before proof generation
Assertion Design: Make assertions specific and verifiable
Key Management: Securely store verification keys
Proof Caching: Cache proofs for repeated verification
Circuit Optimization: Optimize circuits for your use case

Authentication

Endpoints

Request

Response

Proof Types

Performance Proof

Transaction Proof

Compliance Proof

Advanced Features

Batch Proof Generation

Proof Composition

Custom Circuits

Use Cases

Best Practices

Next Steps

Verify Proofs

Attestations

Authentication

Endpoints

​Request

​Response

​Proof Types

​Performance Proof

​Transaction Proof

​Compliance Proof

​Advanced Features

​Batch Proof Generation

​Proof Composition

​Custom Circuits

​Use Cases

​Best Practices

​Next Steps

Verify Proofs

Attestations

Request

Response

Proof Types

Performance Proof

Transaction Proof

Compliance Proof

Advanced Features

Batch Proof Generation

Proof Composition

Custom Circuits

Use Cases

Best Practices

Next Steps