Trust Receipts for
Every AI Interaction
Ed25519 signed receipts for every AI interaction. W3C DIDs for identity. Hash-chained audit trails. Independently verifiable - no trust required.
Cryptographic Trust Infrastructure
Real cryptography, not just hashing. Ed25519 signatures, W3C Decentralized Identifiers, and hash-chained receipts you can verify independently.
Ed25519 Signed Receipts
Every interaction signed with Ed25519 digital signatures. Hash-chained for tamper-evident audit trails. Independently verifiable.
W3C DID Integration
Decentralized Identifiers for platform and agents. Standard did:web method with public key resolution at /.well-known/did.json.
Real-time Policy Scoring
Every AI response evaluated against 6 constitutional principles in under 50ms. Scores from 0-100 with detailed breakdowns.
Violation Alerts
Automatic alerts when AI responses fall below trust thresholds. WebSocket real-time notifications.
Compliance Export
Export signed receipts as JSON for regulatory audits. Full chain verification included. EU AI Act aligned.
Multi-Agent DIDs
Each AI agent gets its own DID with trust metadata. Platform acts as controller with full audit lineage.
The SONATE Trust Protocol
Every AI response is evaluated against 6 constitutional principles. The weighted scores combine into a single trust score (0-100) that determines policy compliance.
How It Works
SONATE sits between your application and AI providers, evaluating every interaction in real-time.
User Sends Message
Request goes through SONATE
AI Generates Response
Response captured for evaluation
Trust Evaluation
6 principles scored in real-time
Receipt Generated
Cryptographic proof stored
Open Verification, Protected Core
We believe in transparency for trust verification while protecting the innovation that makes it possible.
Verification SDK
MIT LicenseIndependently verify trust receipts in your own applications. Full Ed25519 signature verification, chain hash validation, and TypeScript support.
- Node.js & browser support
- Zero dependencies
- Full TypeScript definitions
Core Platform
ProprietaryThe policy engine, trust scoring algorithms, and receipt generation are proprietary. Source available for transparency, commercial licensing available.
- Policy engine & scoring
- Receipt generation & signing
- Enterprise dashboard
Why this model? Anyone can verify receipts independently (trust), but the core technology remains protected (business sustainability).