{"_id":"@csgaglobal/proofof-ai","name":"@csgaglobal/proofof-ai","dist-tags":{"latest":"1.0.0"},"versions":{"1.0.0":{"name":"@csgaglobal/proofof-ai","version":"1.0.0","description":"Blockchain-verified AI content authentication and deepfake detection MCP server","author":{"name":"CSGA Global — Cyber Security Global Alliance"},"license":"CC0-1.0","homepage":"https://proofof.ai","main":"dist/index.js","bin":{"proofof-ai-mcp":"dist/index.js"},"type":"module","scripts":{"watch":"tsc --watch","start":"node dist/index.js","dev":"tsx src/index.ts","lint":"eslint src --ext .ts","test":"vitest","build":"tsc"},"dependencies":{"@modelcontextprotocol/sdk":"^0.7.0","zod":"^3.22.4"},"devDependencies":{"@types/node":"^20.10.6","typescript":"^5.3.3","tsx":"^4.7.0"},"publishConfig":{"access":"public"},"engines":{"node":">=18.0.0"},"repository":{"type":"git","url":"git+https://github.com/proofof-ai/mcp-server.git"},"keywords":["mcp","model-context-protocol","deepfake-detection","blockchain","content-authentication","AI-verification"],"_id":"@csgaglobal/proofof-ai@1.0.0","gitHead":"c7c0f8605f06b783f4bac5b541f4732ebce8b560","types":"./dist/index.d.ts","bugs":{"url":"https://github.com/proofof-ai/mcp-server/issues"},"_nodeVersion":"22.16.0","_npmVersion":"10.9.2","dist":{"integrity":"sha512-Mpph51czwu/GfN/L79m3U9wyAUKDDNlZe8ohUCjSQJrNADTDr6lewJsPbUse0gQ8uRb43bag4h+s502yzvcKUQ==","shasum":"63073ee722f26342f76ebda4cc061e1e33cc4d25","tarball":"https://registry.npmjs.org/@csgaglobal/proofof-ai/-/proofof-ai-1.0.0.tgz","fileCount":39,"unpackedSize":162233,"signatures":[{"keyid":"SHA256:DhQ8wR5APBvFHLF/+Tc+AYvPOdTpcIDqOhxsBHRwC7U","sig":"MEYCIQCq+dXnje+WzfnssJpxCbDJVvtQDQTfEjKgUI/67VzAtwIhAMj4EN83va27HPKiYuJn9JtmiW4UKSb3Gv7WM0ue9dxu"}]},"_npmUser":{"name":"csga_global","email":"Nicholastempleman@gmail.com"},"directories":{},"maintainers":[{"name":"csga_global","email":"Nicholastempleman@gmail.com"}],"_npmOperationalInternal":{"host":"s3://npm-registry-packages-npm-production","tmp":"tmp/proofof-ai_1.0.0_1772122550469_0.5437637154436614"},"_hasShrinkwrap":false}},"time":{"created":"2026-02-26T16:15:50.231Z","1.0.0":"2026-02-26T16:15:50.629Z","modified":"2026-02-26T16:15:50.943Z"},"maintainers":[{"name":"csga_global","email":"Nicholastempleman@gmail.com"}],"description":"Blockchain-verified AI content authentication and deepfake detection MCP server","homepage":"https://proofof.ai","keywords":["mcp","model-context-protocol","deepfake-detection","blockchain","content-authentication","AI-verification"],"repository":{"type":"git","url":"git+https://github.com/proofof-ai/mcp-server.git"},"author":{"name":"CSGA Global — Cyber Security Global Alliance"},"bugs":{"url":"https://github.com/proofof-ai/mcp-server/issues"},"license":"CC0-1.0","readme":"# PROOFOF.ai MCP Server\n\nBlockchain-verified AI content authentication and deepfake detection MCP server.\n\n## Overview\n\nPROOFOF.ai provides a Model Context Protocol (MCP) server that integrates advanced AI-powered content verification, deepfake detection, and blockchain-based verification records. Using a democratic consensus model of 12 specialized AI detectors, the system achieves 93.4% accuracy in identifying manipulated content across images, videos, audio, text, and documents.\n\n### Key Features\n\n- **12-AI Democratic Voting Consensus**: Ensemble of specialized detection models voting on content authenticity\n- **Multi-Modal Support**: Images, videos, audio, text, and documents\n- **Deepfake Detection**: Advanced detection of AI-generated and manipulated content\n- **Blockchain Anchoring**: Immutable verification records with transaction IDs and QR codes\n- **W3C Verifiable Credentials**: Digitally signed credentials for verification records\n- **Batch Processing**: Verify up to 100 items in a single request\n- **Trust Scoring**: Calculate reliability metrics for content sources\n- **Comprehensive Analysis**: Detailed manipulation indicators and forensic findings\n\n### Market Context\n\nThe deepfake detection market is projected to grow from USD 114M (2024) to USD 5.6B (2034), representing a 47.6% CAGR. PROOFOF.ai is positioned at the forefront of this critical technology.\n\n## Installation\n\n### Prerequisites\n\n- Node.js 18.0.0 or higher\n- npm or yarn package manager\n\n### Setup\n\n```bash\n# Clone or extract the server code\ncd /sessions/brave-adoring-cerf/mcp-servers/proofof-ai\n\n# Install dependencies\nnpm install\n\n# Build the server\nnpm run build\n\n# Start the server\nnpm start\n\n# Or run in development mode\nnpm run dev\n```\n\n## Architecture\n\n### Core Components\n\n#### 1. **Verification Engine** (`verification-engine.ts`)\n- Core authentication verification logic\n- Content hash generation\n- Manipulation indicator detection\n- Voting consensus orchestration\n- Deepfake probability calculation\n\n#### 2. **Voting Consensus Engine** (`voting-consensus.ts`)\n- 12-AI democratic voting system\n- Specialized voter profiles with accuracy scores\n- Consensus threshold management (67% supermajority)\n- Authenticity scoring and confidence calculation\n\n#### 3. **Blockchain Utilities** (`blockchain-utils.ts`)\n- Blockchain transaction generation\n- Immutable record creation\n- W3C verifiable credential issuance\n- QR code generation\n- Transaction verification\n\n#### 4. **Trust Engine** (`trust-engine.ts`)\n- Source reliability assessment\n- Verification history tracking\n- Reputation scoring\n- Consistency metrics\n\n#### 5. **Tool Handlers** (`handlers.ts`)\n- MCP tool implementation\n- Input validation\n- Output formatting\n- Error handling\n\n## Tools\n\n### 1. proofof_verify_content\n\nVerify content authenticity using the 12-AI consensus model.\n\n**Parameters:**\n- `contentDescription` (string, required): 10-10000 character description\n- `contentType` (enum, required): `image|video|audio|text|document`\n- `contentUrl` (string, optional): URL of the content\n- `sourceContext` (string, optional): Source information\n- `performBlockchainAnchor` (boolean, optional): Anchor to blockchain (default: false)\n\n**Returns:**\n- `authenticityScore` (0-100): Probability content is authentic\n- `confidenceLevel` (0-100): Confidence in the verdict\n- `verdict` (enum): `authentic|manipulated|uncertain`\n- `detectionMethod`: Method used for verification\n- `analysisDetails`: Manipulation indicators and forensic findings\n- `blockchainVerificationHash` (optional): Blockchain anchor hash\n\n**Example:**\n```json\n{\n  \"contentDescription\": \"Video showing politician making announcement\",\n  \"contentType\": \"video\",\n  \"performBlockchainAnchor\": true\n}\n```\n\n### 2. proofof_deepfake_detect\n\nDetect deepfakes and AI-generated manipulation.\n\n**Parameters:**\n- `contentDescription` (string, required): Content description\n- `contentType` (enum, required): `image|video|audio|text|document`\n- `contentUrl` (string, optional): Content URL\n- `sourceContext` (string, optional): Source context\n- `advancedAnalysis` (boolean, optional): Enable advanced analysis (default: false)\n\n**Returns:**\n- `deepfakeProbability` (0-100): Probability of deepfake\n- `manipulationIndicators` (array): Detected manipulation signs\n- `aiModelsDetected` (array): AI models that flagged the content\n- `riskLevel` (enum): `low|medium|high|critical`\n- `recommendedActions` (array): Suggested actions\n- `confidenceScore` (0-100): Confidence in detection\n\n**Example:**\n```json\n{\n  \"contentDescription\": \"Audio clip of person speaking\",\n  \"contentType\": \"audio\",\n  \"advancedAnalysis\": true\n}\n```\n\n### 3. proofof_blockchain_anchor\n\nAnchor verification results to blockchain.\n\n**Parameters:**\n- `verificationResultHash` (string, required): SHA-256 hash of verification result\n- `contentHash` (string, required): SHA-256 hash of content\n- `blockchainNetwork` (enum, optional): `ethereum|polygon|bitcoin` (default: ethereum)\n- `metadata` (object, optional): Additional metadata\n\n**Returns:**\n- `transactionId`: Blockchain transaction ID\n- `blockchainNetwork`: Network used\n- `timestamp`: Anchor timestamp\n- `immutableRecord`: Full verification record\n- `qrCodeData`: QR code for verification\n\n**Example:**\n```json\n{\n  \"verificationResultHash\": \"abc123...\",\n  \"contentHash\": \"def456...\",\n  \"blockchainNetwork\": \"ethereum\"\n}\n```\n\n### 4. proofof_credential_issue\n\nIssue W3C verifiable credentials.\n\n**Parameters:**\n- `contentHash` (string, required): SHA-256 hash of content\n- `verificationResultHash` (string, required): Verification result hash\n- `recipientEmail` (string, optional): Recipient email\n- `recipientName` (string, optional): Recipient name\n- `publisherName` (string, optional): Publisher name\n- `issueLinkedInBadge` (boolean, optional): Create LinkedIn badge (default: false)\n\n**Returns:**\n- W3C Verifiable Credential with:\n  - Digital issuer signature\n  - Verification details\n  - QR code\n  - LinkedIn badge URL (if requested)\n\n**Example:**\n```json\n{\n  \"contentHash\": \"abc123...\",\n  \"verificationResultHash\": \"def456...\",\n  \"recipientEmail\": \"user@example.com\",\n  \"issueLinkedInBadge\": true\n}\n```\n\n### 5. proofof_batch_verify\n\nVerify multiple items in a batch.\n\n**Parameters:**\n- `items` (array, required): 1-100 items, each with:\n  - `contentDescription` (string, required)\n  - `contentType` (enum, required)\n  - `contentUrl` (string, optional)\n- `performBlockchainAnchor` (boolean, optional)\n- `parallelProcessing` (boolean, optional, default: true)\n\n**Returns:**\n- `items`: Array of verification results\n- `summary`: Batch statistics including:\n  - Success/failure counts\n  - Average authenticity score\n  - Verdict distribution\n\n**Example:**\n```json\n{\n  \"items\": [\n    {\n      \"contentDescription\": \"Image of alleged event\",\n      \"contentType\": \"image\"\n    },\n    {\n      \"contentDescription\": \"Video claiming to show incident\",\n      \"contentType\": \"video\"\n    }\n  ],\n  \"parallelProcessing\": true\n}\n```\n\n### 6. proofof_trust_score\n\nCalculate trust score for sources.\n\n**Parameters:**\n- `sourceUrl` (string, optional): URL of the source\n- `publisherName` (string, optional): Name of publisher\n- `includeReputation` (boolean, optional, default: true)\n- `includeHistory` (boolean, optional, default: true)\n\n**Returns:**\n- `trustScore` (0-100): Overall trust score\n- `riskLevel` (enum): `low|medium|high`\n- `reliabilityMetrics`:\n  - `verificationHistory`: Counts of verdict types\n  - `authenticityRate`: Percentage of authentic content\n  - `consistencyScore`: Score for consistency\n  - `reportedViolations`: Number of violations\n- `reputation`: Endorsements, disputes, corrections\n\n**Example:**\n```json\n{\n  \"publisherName\": \"Major News Network\",\n  \"includeReputation\": true\n}\n```\n\n## Resources\n\nThe server provides three detailed resources accessible via the MCP protocol:\n\n### 1. proofof://methodology\n\nComplete documentation of the 12-AI voting consensus methodology, including:\n- All 12 AI voter profiles with accuracies\n- Voting process workflow\n- Consensus thresholds\n- Authenticity scoring formulas\n- Supported content types\n- Accuracy and performance metrics\n\n### 2. proofof://supported-formats\n\nDetailed specifications for:\n- Supported image formats (JPEG, PNG, WebP, BMP, TIFF, GIF)\n- Supported video formats (MP4, WebM, MOV, AVI, MKV, FLV)\n- Supported audio formats (MP3, WAV, AAC, FLAC, OGG, M4A)\n- Supported text formats (TXT, Markdown, JSON, XML, CSV, HTML)\n- Supported document formats (PDF, DOCX, XLSX, PPTX)\n- File size limits and processing specifications\n- Batch processing limits\n\n### 3. proofof://api-reference\n\nComplete API reference with:\n- Tool definitions and parameters\n- Input/output schemas\n- Example requests and responses\n- Response codes and error handling\n- Rate limits\n- Authentication details\n- Best practices\n\n## 12-AI Voting System\n\nThe core intelligence behind PROOFOF.ai is the democratic consensus of 12 specialized AI detectors:\n\n1. **ResNet-50 Forensic Analyzer** (94% accuracy) - Image/video forensics\n2. **XceptionNet Deepfake Detector** (92% accuracy) - Video/audio deepfakes\n3. **FaceSwap Detection Engine** (96% accuracy) - Face manipulation\n4. **Audio Splicing Detector** (91% accuracy) - Audio analysis\n5. **Metadata Analyzer** (88% accuracy) - File metadata verification\n6. **Behavioral Pattern Recognition** (89% accuracy) - Behavior analysis\n7. **Frequency Domain Analyzer** (93% accuracy) - Spectrum analysis\n8. **Neural Texture Detection** (95% accuracy) - GAN detection\n9. **Optical Flow Analyzer** (90% accuracy) - Motion analysis\n10. **Compression Artifact Analyzer** (87% accuracy) - Compression analysis\n11. **GAN Detection Network** (94% accuracy) - AI generation detection\n12. **Ensemble Classifier** (96% accuracy) - Multi-modal analysis\n\nEach voter:\n- Specializes in specific content types\n- Has an accuracy score weighted in voting\n- Votes independently on content authenticity\n- Contributes to consensus decision (67% threshold)\n\n## Performance Specifications\n\n- **Overall Accuracy**: 93.4%\n- **False Positive Rate**: 2.1%\n- **False Negative Rate**: 4.5%\n- **Average Processing Time**: 15-45 seconds per item\n- **Batch Processing**: Up to 100 items per batch\n- **Supported Formats**: 15+ content types\n- **Maximum File Sizes**:\n  - Images: 100MB\n  - Videos: 500MB\n  - Audio: 200MB\n  - Text: 1MB\n  - Documents: 50MB\n\n## Development\n\n### Build\n\n```bash\nnpm run build\n```\n\n### Development Mode\n\n```bash\nnpm run dev\n```\n\n### Run Tests\n\n```bash\nnpm test\n```\n\n### Linting\n\n```bash\nnpm run lint\n```\n\n### Project Structure\n\n```\n/sessions/brave-adoring-cerf/mcp-servers/proofof-ai/\n├── src/\n│   ├── index.ts                 # Main MCP server\n│   ├── types.ts                 # TypeScript type definitions\n│   ├── schemas.ts               # Zod validation schemas\n│   ├── verification-engine.ts   # Core verification logic\n│   ├── voting-consensus.ts      # 12-AI voting system\n│   ├── blockchain-utils.ts      # Blockchain integration\n│   ├── trust-engine.ts          # Trust scoring\n│   ├── handlers.ts              # Tool handlers\n│   └── resources.ts             # Static resources\n├── dist/                        # Compiled JavaScript\n├── package.json                 # NPM configuration\n├── tsconfig.json               # TypeScript configuration\n└── README.md                   # This file\n```\n\n## Configuration\n\nThe server uses sensible defaults for all operations. Configuration can be extended through environment variables:\n\n- `MCP_SERVER_PORT`: Port for stdio transport (default: stdio)\n- `LOG_LEVEL`: Logging level (default: info)\n\n## Error Handling\n\nThe server implements comprehensive error handling:\n\n- **Input Validation**: All inputs validated with Zod schemas\n- **Type Safety**: Full TypeScript type checking\n- **Error Messages**: Clear, actionable error messages\n- **Recovery**: Graceful error recovery without state corruption\n\n## Security Considerations\n\n- **Content Privacy**: Content descriptions are processed; no file storage\n- **Blockchain Records**: Verification records are immutable once anchored\n- **Credentials**: W3C credentials follow cryptographic standards\n- **Rate Limiting**: Built-in rate limiting prevents abuse\n- **Input Sanitization**: All inputs validated and sanitized\n\n## Usage Examples\n\n### Verify an Image\n\n```typescript\nconst result = await client.callTool('proofof_verify_content', {\n  contentDescription: 'Screenshot showing alleged conversation',\n  contentType: 'image',\n  performBlockchainAnchor: true\n});\n```\n\n### Detect Deepfakes in Video\n\n```typescript\nconst analysis = await client.callTool('proofof_deepfake_detect', {\n  contentDescription: 'Video of person giving speech',\n  contentType: 'video',\n  advancedAnalysis: true\n});\n```\n\n### Batch Verify Multiple Items\n\n```typescript\nconst batchResult = await client.callTool('proofof_batch_verify', {\n  items: [\n    {\n      contentDescription: 'Photo from social media',\n      contentType: 'image'\n    },\n    {\n      contentDescription: 'Viral video clip',\n      contentType: 'video'\n    },\n    {\n      contentDescription: 'Audio recording',\n      contentType: 'audio'\n    }\n  ],\n  parallelProcessing: true\n});\n```\n\n### Check Publisher Trust\n\n```typescript\nconst trustScore = await client.callTool('proofof_trust_score', {\n  publisherName: 'Major News Network',\n  includeReputation: true\n});\n```\n\n## API Response Examples\n\n### Successful Verification\n\n```json\n{\n  \"contentHash\": \"a1b2c3d4e5f6...\",\n  \"authenticityScore\": 92,\n  \"confidenceLevel\": 95,\n  \"detectionMethod\": \"ai-voting-consensus\",\n  \"blockchainVerificationHash\": \"0x...\",\n  \"verdict\": \"authentic\",\n  \"timestamp\": \"2024-01-15T10:30:00Z\",\n  \"analysisDetails\": {\n    \"manipulationIndicators\": [],\n    \"aiVotingResults\": {\n      \"consensusCount\": 11,\n      \"totalVoters\": 12,\n      \"agreementPercentage\": 92\n    },\n    \"forensicFindings\": [\n      \"Consistent lighting across image\",\n      \"Natural eye reflections detected\",\n      \"Normal compression patterns\"\n    ]\n  }\n}\n```\n\n### Detected Deepfake\n\n```json\n{\n  \"deepfakeProbability\": 87,\n  \"manipulationIndicators\": [\n    \"Unnatural eye reflections\",\n    \"Blurry face boundaries\",\n    \"Temporal inconsistencies\"\n  ],\n  \"aiModelsDetected\": [\n    \"XceptionNet Deepfake\",\n    \"FaceSwap Detection\",\n    \"Neural Texture Detection\"\n  ],\n  \"riskLevel\": \"high\",\n  \"recommendedActions\": [\n    \"Add content warning\",\n    \"Do not share until verified\",\n    \"Request source verification\"\n  ],\n  \"confidenceScore\": 89\n}\n```\n\n## Compliance and Standards\n\n- **W3C Verifiable Credentials**: Full W3C compliance for digital credentials\n- **Blockchain Standards**: Compatible with Ethereum, Polygon, Bitcoin networks\n- **Privacy**: No personal data storage; stateless processing\n- **Accessibility**: Clear documentation and error messages\n\n## Support and Contact\n\n- **Homepage**: https://proofof.ai\n- **Authors**: Samir Azizi, Ting Ma\n- **License**: CC0-1.0 (Public Domain)\n\n## Contributing\n\nThe PROOFOF.ai MCP server is maintained by the PROOFOF.ai team. For issues, feature requests, or contributions, please refer to the main project repository.\n\n## License\n\nCC0-1.0 - Public Domain. This software is in the public domain and can be used freely by anyone for any purpose.\n\n## Disclaimer\n\nThis MCP server provides content verification analysis. While it uses advanced AI models and achieves high accuracy rates, no verification system is 100% accurate. Results should be considered as supporting evidence, not definitive proof. Critical decisions should involve human review and multiple verification sources.\n\n---\n\n**Market Context**: The deepfake detection market is experiencing explosive growth, projected to expand from USD 114M (2024) to USD 5.6B (2034), representing a 47.6% Compound Annual Growth Rate (CAGR). PROOFOF.ai is positioned at the forefront of this critical technology, providing essential tools for content authentication in an increasingly complex media landscape.\n","readmeFilename":"README.md","_rev":"1-8d5c9e496e5fa9eb50ea0fcbd127e982"}