ProofsyGPU - GPU Job Receipt Tracker

🎯 Overview

ProofsyGPU is a serverless GPU job receipt tracking system that logs compute job lifecycles (JobSubmitted → JobCompleted) as ERC-7053 commit events to Numbers Mainnet blockchain. Built with Vercel serverless functions and provides a compact audit UI for transparent GPU usage accountability.

🛡️ C2PA Content Provenance Integration

ProofsyGPU implements C2PA (Coalition for Content Provenance and Authenticity) for AI-generated content verification:

📄 JSON Artifacts: Detailed GPU job results with embedded C2PA manifests
🔐 Cryptographic Signatures: RSA-based signatures for artifact authenticity
🔗 Blockchain Proof: Embedded Numbers Protocol NIDs for blockchain verification
✅ Self-Contained Verification: Download artifacts with embedded proof
🎯 Complete Provenance Chain: From job submission to artifact generation

✨ Key Features

🔗 Blockchain Job Receipts: Every GPU job logged to Numbers Mainnet with ERC-7053 standard
🛡️ C2PA Content Provenance: JSON artifacts with embedded blockchain proof and cryptographic signatures
⚡ Serverless Architecture: Vercel serverless functions with auto-scaling
🤖 Mock GPU Simulator: Realistic job simulation without requiring real GPU hardware
📊 Audit UI: Query job timeline by NID with complete blockchain proof and C2PA verification
🎯 5 Task Types: Stable Diffusion, LLM Inference, Model Training, Video Upscaling, Speech-to-Text
🔒 Idempotency: Prevents duplicate event processing
📱 Real-time Status: System health monitoring and job tracking
✅ C2PA Verification: Download and verify signed artifacts with embedded blockchain proof

🏗️ Architecture

System Architecture Diagram

graph TB
    subgraph "Frontend - Static HTML/JS"
        UI[Web Interface]
        JobForm[Job Submission Form]
        Timeline[Audit Timeline UI]
        JobList[Jobs Dashboard]
    end
    
    subgraph "Vercel Serverless Functions"
        SubmitAPI[api/jobs/submit.js]
        CompleteAPI[api/jobs/complete.js]
        TimelineAPI[api/jobs/timeline/jobId.js]
        ListAPI[api/jobs/list.js]
        HealthAPI[api/health.js]
        DownloadAPI[api/artifacts/download/jobId.js]
        VerifyAPI[api/artifacts/verify-c2pa.js]
    end
    
    subgraph "Mock GPU Layer"
        Simulator[GPU Job Simulator]
        TaskTypes[5 Task Types]
        ArtifactGen[JSON Artifact Generator]
        C2PASigner[C2PA Signature Engine]
    end
    
    subgraph "Data Layer"
        Prisma[Prisma ORM]
        VercelDB[(Vercel Postgres)]
    end
    
    subgraph "Blockchain Layer"
        Numbers[Numbers Protocol<br/>Capture API]
        Mainnet[Numbers Mainnet<br/>ERC-7053]
        Explorer[Blockchain Explorer]
    end
    
    UI --> JobForm
    UI --> Timeline
    UI --> JobList
    
    JobForm -->|POST| SubmitAPI
    Timeline -->|GET| TimelineAPI
    JobList -->|GET| ListAPI
    
    SubmitAPI --> Simulator
    CompleteAPI --> Simulator
    
    Simulator --> TaskTypes
    Simulator --> ArtifactGen
    
    SubmitAPI -->|Commit Receipt| Numbers
    CompleteAPI -->|Commit Receipt| Numbers
    
    SubmitAPI -->|Save| Prisma
    CompleteAPI -->|Save| Prisma
    TimelineAPI -->|Query| Prisma
    
    Prisma --> VercelDB
    
    Numbers -->|ERC-7053| Mainnet
    Mainnet -->|NID + TxHash| Numbers
    
    Timeline -->|Verify| Explorer
    
    style UI fill:#e1f5ff
    style SubmitAPI fill:#fff4e1
    style Simulator fill:#f0e1ff
    style VercelDB fill:#e1ffe1
    style Mainnet fill:#ffe1e1

Job Lifecycle Flow

sequenceDiagram
    participant User
    participant Frontend
    participant SubmitAPI as /api/jobs/submit
    participant Simulator as GPU Simulator
    participant Numbers as Numbers API
    participant Blockchain as Numbers Mainnet
    participant DB as Vercel Postgres
    participant CompleteAPI as /api/jobs/complete
    participant Explorer
    
    User->>Frontend: 1. Select Task Type<br/>(Stable Diffusion, etc.)
    Frontend->>SubmitAPI: 2. POST /api/jobs/submit<br/>{taskType, executor}
    
    SubmitAPI->>Simulator: 3. Generate Job Data
    Simulator-->>SubmitAPI: jobId, gpuType, inputHash
    
    SubmitAPI->>Numbers: 4. Commit JobSubmitted<br/>(ERC-7053)
    Numbers->>Blockchain: 5. Write to Mainnet
    Blockchain-->>Numbers: 6. Return NID + TxHash
    Numbers-->>SubmitAPI: Receipt
    
    SubmitAPI->>DB: 7. Save JobReceipt<br/>(idempotencyKey)
    DB-->>SubmitAPI: ✅ Stored
    
    SubmitAPI-->>Frontend: 8. Success + NID
    Frontend-->>User: 9. Job Submitted ✓
    
    Note over Frontend,CompleteAPI: Auto-complete after 3s (optional)
    
    Frontend->>CompleteAPI: 10. POST /api/jobs/complete<br/>{jobId}
    CompleteAPI->>Simulator: 11. Generate Output
    Simulator-->>CompleteAPI: outputHash, duration, gpuUtilization
    
    CompleteAPI->>Numbers: 12. Commit JobCompleted<br/>(ERC-7053)
    Numbers->>Blockchain: 13. Write to Mainnet
    Blockchain-->>Numbers: 14. Return NID + TxHash
    
    CompleteAPI->>DB: 15. Save Completion Receipt
    CompleteAPI-->>Frontend: 16. Success + Receipt
    
    User->>Frontend: 17. View Timeline
    Frontend->>Explorer: 18. Verify on Blockchain
    Explorer-->>User: 19. Show Proof ✅

📋 Event Schema

JobSubmitted Event

{
  "eventType": "JobSubmitted",
  "jobId": "job_1729012345_abc123",
  "taskType": "stable-diffusion",
  "executor": "0x51130dB91B91377A24d6Ebeb2a5fC02748b53ce1",
  "occurredAt": "2025-10-15T10:00:00Z",
  "idempotencyKey": "gpu-job-job_1729012345_abc123-JobSubmitted",
  "metadata": {
    "jobId": "job_1729012345_abc123",
    "taskType": "stable-diffusion",
    "taskName": "Stable Diffusion Image Generation",
    "gpuType": "NVIDIA A100 80GB",
    "estimatedDuration": 45,
    "inputHash": "QmInputAbc123...",
    "executor": "0x51130dB91B91377A24d6Ebeb2a5fC02748b53ce1",
    "status": "submitted"
  }
}

JobCompleted Event

{
  "eventType": "JobCompleted",
  "jobId": "job_1729012345_abc123",
  "taskType": "stable-diffusion",
  "executor": "0x51130dB91B91377A24d6Ebeb2a5fC02748b53ce1",
  "occurredAt": "2025-10-15T10:00:52Z",
  "idempotencyKey": "gpu-job-job_1729012345_abc123-JobCompleted",
  "metadata": {
    "jobId": "job_1729012345_abc123",
    "taskType": "stable-diffusion",
    "taskName": "Stable Diffusion Image Generation",
    "gpuType": "NVIDIA A100 80GB",
    "executor": "0x51130dB91B91377A24d6Ebeb2a5fC02748b53ce1",
    "outputHash": "QmOutputDef456...",
    "outputCid": "bafybeigabc123...",
    "duration": 52.3,
    "gpuUtilization": "87%",
    "exitCode": 0,
    "status": "completed",
    "artifact": {
      "fileName": "job_1729012345_abc123_output.png",
      "fileHash": "sha256:abc123...",
      "fileSize": 2048000,
      "mimeType": "image/png"
    }
  }
}

Blockchain Storage Schema

{
  "assetCid": "bafkrei...",
  "assetSha256": "proof_hash_from_numbers_api",
  "encodingFormat": "application/json",
  "assetTimestampCreated": 1760615889,
  "assetCreator": "creator_name",
  "abstract": "GPU Job JobSubmitted: taskType | JobID: jobId | GPU: gpuType | Executor: executor",
  "headline": "ProofsyGPU - taskType Job Receipt",
  "assetSourceType": "others",
  "creatorWallet": "0x...",
  "jobDetails": {
    "eventType": "JobSubmitted|JobCompleted",
    "jobId": "job_id",
    "taskType": "task_type",
    "executor": "0x...",
    "occurredAt": "ISO_timestamp",
    "idempotencyKey": "gpu-job-jobId-eventType"
  },
  "gpuJobMetadata": {
    "jobId": "job_id",
    "taskType": "task_type",
    "taskName": "Human Readable Task Name",
    "gpuType": "NVIDIA A100 80GB",
    "executor": "0x...",
    "status": "submitted|completed",
    "inputHash": "QmInput...",
    "estimatedDuration": 45,
    "outputHash": "QmOutput...",
    "duration": 52.3,
    "gpuUtilization": "87%",
    "artifact": { "fileName": "...", "fileHash": "...", "fileSize": 1234, "mimeType": "..." }
  }
}

🚀 Quick Start

Prerequisites

Node.js 18+
Vercel account
Numbers Protocol API Key

Local Development

Clone Repository

git clone https://github.com/yourusername/proofsy-gpu.git
cd proofsy-gpu

Install Dependencies

npm install

Key Dependencies:

@prisma/client - Database ORM
node-forge - C2PA cryptographic signatures
@vercel/postgres - Database connection

Configure Environment

cp .env.example .env
# Edit .env with your Numbers Protocol API key

Setup Database

npx prisma migrate dev

Start Development Server

npm run dev

Access Application Open browser: http://localhost:3000

🚀 Vercel Deployment

One-Click Deploy

Manual Deployment

Install Vercel CLI

npm i -g vercel

Login to Vercel

vercel login

Deploy

vercel --prod

Add Vercel Postgres Database
- Go to Vercel Dashboard → Storage → Create Database
- Select PostgreSQL
- Auto-links to your project
- Vercel auto-injects POSTGRES_PRISMA_URL and other DB variables
Configure Environment Variables in Vercel Dashboard
- NUMBERS_API_KEY - Your Numbers Protocol API key
- Database URLs are auto-configured by Vercel Postgres
Run Database Migrations
- Migrations run automatically on first deploy via npm run build
- Or manually: vercel env pull .env.local && npx prisma migrate deploy

🎮 Mock GPU Simulator

Supported Task Types

Stable Diffusion Image Generation
- Simulated duration: ~45s
- Output: PNG image (mock)
LLM Text Inference
- Simulated duration: ~12s
- Output: JSON response (mock)
Model Fine-tuning
- Simulated duration: ~300s
- Output: .pt model file (mock)
AI Video Upscaling
- Simulated duration: ~180s
- Output: MP4 video (mock)
Speech-to-Text Transcription
- Simulated duration: ~30s
- Output: JSON transcript (mock)

GPU Types (Simulated)

NVIDIA A100 80GB
NVIDIA H100 SXM
NVIDIA RTX 4090
AMD MI300X
Google TPU v5e

Realistic Metrics

GPU Utilization: 70-95% (randomized)
Duration: ±20% variance from estimated
Exit codes, output hashes, CIDs

🔧 API Endpoints

Health Check

GET /api/health

Submit Job

POST /api/jobs/submit
Content-Type: application/json

{
  "taskType": "stable-diffusion",
  "executor": "0x..."
}

Complete Job

POST /api/jobs/complete
Content-Type: application/json

{
  "jobId": "job_...",
  "taskType": "stable-diffusion",
  "estimatedDuration": 45,
  "executor": "0x..."
}

Get Job Timeline

GET /api/jobs/timeline/{jobId}

List All Jobs

GET /api/jobs/list

Download C2PA-Signed Artifact

GET /api/artifacts/download/{jobId}

Response: JSON artifact with embedded C2PA manifest and blockchain proof

Verify C2PA Signature

POST /api/artifacts/verify-c2pa
Content-Type: application/json

{
  "artifactContent": "{...json artifact...}"
}

Response: Verification result with manifest details

🛡️ C2PA Implementation

C2PA Artifact Structure

{
  "jobId": "job_1760648655376_nlzi97ldh",
  "taskType": "video-upscaling",
  "executedAt": "2025-10-16T21:04:23.377Z",
  "executor": "0x51130dB91B91377A24d6Ebeb2a5fC02748b53ce1",
  "result": {
    "inputResolution": "1920x1080",
    "outputResolution": "3840x2160",
    "algorithm": "ESRGAN",
    "upscaleFactor": 2,
    "processingTime": 180.5,
    "videoUrl": "ipfs://QmVideort2qdxrsn0b"
  },
  "execution": {
    "gpuType": "NVIDIA H100 SXM",
    "duration": 163.1,
    "gpuUtilization": "74%",
    "memoryUsed": "45GB",
    "exitCode": 0
  },
  "hashes": {
    "inputHash": "QmInput2xuzvscljjk",
    "outputHash": "sha256:abc123def456..."
  },
  "c2pa": {
    "version": "1.0",
    "claim_generator": "ProofsyGPU/1.0",
    "assertions": [
      {
        "label": "gpu.job.provenance",
        "data": {
          "jobId": "job_1760648655376_nlzi97ldh",
          "taskType": "video-upscaling",
          "executor": "0x51130dB91B91377A24d6Ebeb2a5fC02748b53ce1",
          "gpuType": "NVIDIA H100 SXM",
          "occurredAt": "2025-10-16T21:04:23.377Z"
        }
      },
      {
        "label": "blockchain.proof",
        "data": {
          "chain": "numbers-mainnet",
          "submittedNid": "bafkrei...",
          "completedNid": "bafkrei...",
          "explorerUrl": "https://verify.numbersprotocol.io/..."
        }
      }
    ],
    "signature": {
      "algorithm": "es256",
      "value": "MEUCIQDx...",
      "publicKey": "-----BEGIN PUBLIC KEY-----...",
      "signedAt": "2025-10-16T21:04:23.377Z"
    }
  }
}

C2PA Verification Flow

Job Completion: Generate JSON artifact with GPU results
C2PA Signing: Embed blockchain proof and cryptographic signature
Storage: Save signed artifact to database
Download: Users can download artifacts with embedded proof
Verification: Verify signature and blockchain proof integrity

C2PA Benefits

Self-Contained Proof: Artifact contains complete verification data
Blockchain Integration: Links to Numbers Protocol NIDs
Cryptographic Integrity: RSA-based signatures for authenticity
Human-Readable: JSON format with clear structure
Portable: Download and verify anywhere

🔍 Ordering & Idempotency

Idempotency Strategy

Deterministic Keys: Each event has idempotencyKey: gpu-job-{jobId}-{eventType} (no timestamp)
Duplicate Prevention: Database enforces uniqueness via @@unique constraint
Retry Handling: API returns 409 Conflict if same job+event submitted twice
Example:
- gpu-job-job_123_abc-JobSubmitted (first submit = success)
- gpu-job-job_123_abc-JobSubmitted (retry = 409 Conflict)
- gpu-job-job_123_abc-JobCompleted (different event = success)

Ordering Strategy

Events ordered by occurredAt timestamp
Timeline query sorts chronologically (oldest first)
JobSubmitted always precedes JobCompleted for same jobId
Deterministic keys ensure exactly-once semantics per job lifecycle stage

Local Job States → Blockchain Payloads Mapping

graph LR
    subgraph "Local Job State"
        A[Job Submitted] --> B[Job Completed]
    end
    
    subgraph "Blockchain Events"
        C[JobSubmitted Event] --> D[JobCompleted Event]
    end
    
    subgraph "Numbers Protocol Storage"
        E[Asset File: JSON Metadata] 
        F[Caption: Job Summary]
        G[nit_commit_custom: Structured Data]
        H[Information: Proof + Job Details]
    end
    
    A -->|Generate Job Data| C
    B -->|Generate Completion Data| D
    
    C --> E
    C --> F
    C --> G
    C --> H
    
    D --> E
    D --> F
    D --> G
    D --> H
    
    style A fill:#e1f5ff
    style B fill:#e1f5ff
    style C fill:#fff4e1
    style D fill:#fff4e1
    style E fill:#f0e1ff
    style F fill:#f0e1ff
    style G fill:#f0e1ff
    style H fill:#f0e1ff

Job State Transitions

Job Submitted → JobSubmitted Event → Numbers Protocol Asset
Job Completed → JobCompleted Event → Numbers Protocol Asset
Each Event → Unique Blockchain Receipt (NID + Proof Hash)
Timeline Query → Chronological Event Chain → Complete Audit Trail

🔐 Blockchain Verification

Every job receipt includes:

NID: Asset identifier on Numbers Mainnet
Transaction Hash: Blockchain transaction ID
Explorer Link: Direct link to verify on Numbers Protocol Explorer
Metadata: Complete job details stored on-chain

Verify Job:

Click "View Timeline" on any job
Click "Verify on Blockchain" button
See complete on-chain proof at verify.numbersprotocol.io

C2PA Artifact Features:

Download Artifact: Click "📥 Download JSON with Proof" to get C2PA-signed artifact
Verify Signature: Click "✅ Verify C2PA Signature" to verify authenticity
View Results: Expand "📋 Result Preview" to see detailed GPU results
Blockchain Proof: Artifact contains embedded blockchain NIDs and explorer links

📊 Database Schema

model JobReceipt {
  id              String   @id @default(cuid())
  idempotencyKey  String   @unique
  jobId           String
  eventType       String   // "JobSubmitted" or "JobCompleted"
  taskType        String
  executor        String
  occurredAt      DateTime
  metadataJson    String
  inputHash       String?
  outputHash      String?
  gpuType         String?
  duration        Float?
  gpuUtilization  String?
  txHash          String
  nid             String
  chain           String   @default("numbers-mainnet")
  createdAt       DateTime @default(now())
  
  @@index([jobId])
  @@index([nid])
}

📝 License

MIT License

🤝 Contributing

Fork the repository
Create feature branch
Commit changes
Push to branch
Create Pull Request

📞 Support

Documentation: Numbers Protocol Docs
API Reference: Capture API
Vercel Docs: Vercel Platform

ProofsyGPU - Transparent GPU Job Tracking on Numbers Mainnet ⚡

Name		Name	Last commit message	Last commit date
Latest commit History 30 Commits
api		api
lib		lib
prisma		prisma
public		public
.env.example		.env.example
.gitignore		.gitignore
README.md		README.md
package.json		package.json
reset-database.md		reset-database.md
vercel.json		vercel.json

Modolo-oss/Proofsy-GPU

Folders and files

Latest commit

History

Repository files navigation