VAIBot V1 — Provenance API

Prove what your agent created

Anchor AI-generated content to the blockchain. Every output gets a cryptographic receipt — tamper-proof, independently verifiable, and stored on Base mainnet forever.

Trust content, not claims

VAIBot V1 produces cryptographic proof that content was created by your agent — not tampered with, not fabricated after the fact.

SHA-256 fingerprints

Every piece of content gets a canonical SHA-256 hash. Change one character and the fingerprint no longer matches — making tampering immediately detectable.

On-chain anchoring

Receipt hashes are batched into Merkle roots and anchored to Base mainnet. The chain record is permanent and owned by no one — it cannot be deleted.

Tamper-evident proofs

Anyone can independently verify that a document came from your agent, unmodified, at a specific timestamp — without trusting VAIBot as an intermediary.

Full audit trail

Track every tool call, output, and revision in a linked provenance chain. See parent→child relationships across multi-step agent workflows.

What a provenance receipt looks like

db.querysafe
Hash0x8a3f9b…c4e1f4
Latency38ms
StatusRecorded
file.deleteunsafe
Hash0xb2c4d6…a5b7c9
Latency12ms
StatusEscalated
email.sendwarning
Hash0x5f91a3…d1e5f9
Latency67ms
StatusRecorded

See provenance in action

Chat with a simulated agent. Every tool call generates a real receipt — with a hash, risk level, and provenance chain.

VAIBot V1 — Live demo

Provenance in action

Every agent tool call generates a cryptographic receipt. Type a command to see provenance tracking in real time.

VaiBot AgentProvenance demo
Welcome to the VaiBot V1 Provenance Demo. I'm a simulated AI agent — every tool call I make gets a cryptographic provenance receipt. Try asking me to: - "delete file" — see an unsafe action escalation - "send email" — external API call tracking - "query database" — safe tool call recording - "generate image" — image revision history - "analyze data" — multi-step provenance chain
Provenance Panel0 receipts

Send a message to see tool call receipts appear here.

What you get

Your provenance dashboard

Every receipt anchored to Base mainnet — with a direct link to the on-chain transaction. Filter by risk, verify the chain, trace parent receipts.

Provenance dashboard
Demo data
128Total receipts
126Anchored
2Pending anchor
0Failed
WhenContentHash → ChainRiskAnchor
Apr 23 · 15:04
GPT-4o contract draft — Section 4 liability
legal-agent · write
0x8343…5e347SafeAnchored
Apr 23 · 14:51
Claude code review — auth middleware refactor
dev-agent · read
0xc1a2…9f123SafeAnchored
Apr 23 · 14:33
LLM image analysis — product screenshot
vision-agent · analyze
0xd4e5…3b456SafePending
Apr 23 · 13:58
Agent file analysis — dependency audit report
audit-agent · read
0xf6a7…7c789WarningAnchored
All receipts anchored to Base mainnet · Showing 4 of 128 · Sign in to see your live data
+

Pair with Governance for complete coverage

The VAIBot Audit plan combines both products: V2 governs every agent action before it runs, and V1 anchors the governance receipt on-chain. Complete chain of custody — from proposed action to immutable proof.

See Audit plan