Add a tamper-proof evidence layer to your AI platform.
MCP Server, npm SDK, or REST API — pick your path.
High-risk AI systems must maintain execution logs and transparency reports. Chatbots, automated decisions — all in scope.
"Prove what your AI told the customer" — auditors are already asking. Manual logs won't cut it.
AI agents execute workflows autonomously. Every action needs a cryptographic proof trail — not just a log line.
Choose the path that fits your AI platform. Every path delivers the same result: AI execution → hash record → verifiable proof.
AI agents call directly
Claude, Cursor, VS Code — any MCP-compatible AI client calls proof_record and proof_verify directly.
Node.js / TypeScript
Create DPUs and verify chains directly from your backend. Zero dependency, TypeScript native.
Any language, HTTP call
Python, Java, Go — one HTTP call from any stack. Bearer token auth, JSON response.
Standard MCP tools that any AI agent can call directly. Streamable HTTP protocol, session-based auth.
proof_recordCreate DPU
AI execution → SHA-256 hash + timestamp → immutable chain record
proof_verifyVerify Record
Verify cryptographic integrity of a specific proof record
proof_chain_verifyVerify Chain
Verify entire hash chain continuity and integrity for a domain
proof_getGet Details
Retrieve full details including metadata, governance level, and compliance info
proof_export_jsonldExport JSON-LD
Export as JSON-LD v2.0 document conforming to Cronozen Evidence Ontology
proof_public_verifyPublic Verify
Anyone can verify proof integrity without authentication — zero trust
Every time an AI agent executes a customer-facing action — card issuance, booking, data change — Proof Layer records it automatically.
Instant proof of "what the AI told the customer" for auditors
RPA, workflow automation, AI document generation — attach tamper-proof evidence to every automated decision.
Transparent proof of "why this automation ran this way" for your customers
AI-powered risk assessment, loan approval, fraud detection — regulated industries need cryptographic audit trails, not log files.
SOC2 and regulatory audit — pass with cryptographic evidence
Cryptographic hash on every AI execution record
Policy → Evidence → Human Review → Risk Threshold → Dual Approval
Cronozen Evidence Ontology — submit to auditors instantly
3 verticals running in production with zero audit findings
Open-source core. Managed cloud when you need it.
Open Source
Managed by Cronozen
Managed by Cronozen
On-premise or Cloud
MCP integration in 1-2 days. SDK install in 5 minutes.
Prove everything your AI executes.