Verification API live

Your AI agents act. We prove every decision.

AI agent governance with cryptographic verification — identity, policy enforcement, and on-chain proof for every decision.

The governance layer for AI agents in production. Verify what was allowed and what actually happened — independently.

Cryptographic identity — Ed25519 signatures bind every action to its agent
Policy before execution — rate limits, spend caps, and allowlists enforced before the action runs
Tamper-evident audit — only logs anchored on Base L2 with receipt integrity you can verify yourself
Independently verifiable — anyone can check the receipt using the on-chain anchor alone. No Inntris account required.

When your agents handle money, data, and code in regulated environments, you need cryptographic proof of every decision. Inntris is that proof layer.

See live verification →Request Access

<100ms

Verification latency

Ed25519

Signing algorithm

Base L2

Audit anchor

Fail-closed

Default policy mode

Verification decision flow

Where Inntris sits in your stack

live policy path

Action requested

Agent requests a code, data, API, or finance operation.

Policy evaluated

Inntris checks permissions, risk, and execution context.

Decision signed

Approved or blocked outcome is bound to the agent identity.

Proof recorded

Evidence is written to the audit trail for later verification.

Product surface

One control plane. Four modules.

Operate

Prove

Admin Console

For platform admins

Manage agents, enforce policies, rotate API keys, and investigate security alerts from one dashboard.

Explore Console

Agent Portal

For developers and operators

Explore Portal

Audit Explorer

For investigations and compliance

Search every verification decision by agent, action, or verdict. Each record is append-only and tamper-evident.

Explore Audit Explorer

Public Verify

For customers, partners, and auditors

Share a receipt link with anyone. They can independently verify the signature, policy hash, and on-chain anchor.

Verify a live receipt

Use cases

Pull request verificationSensitive data accessAPI and tool executionFinancial operations

Live proof

A real verification receipt.

Every receipt below is live, signed, anchored to Base L2, and independently verifiable. Click any receipt to inspect the full proof — signature, policy hash, and on-chain anchor.

PASS

Safe action evaluated, approved, and recorded on-chain.

View full receipt

Agent

Mainnet Receipt Agent

Global Bank Corp

Action

financial_transaction

Trust 85/100

Signature

Valid Ed25519

Anchored on Base L2

Receipt ID

Generated 15 hours ago · 2026-04-07 19:22 UTC

BLOCK

Amount $75.0 exceeds per-action limit of $50.00.

View full receipt

Agent

Mainnet Receipt Agent

Global Bank Corp

Action

financial_transaction

Trust 85/100

Signature

Valid Ed25519

Anchored on Base L2

Receipt ID

Generated 15 hours ago · 2026-04-07 19:22 UTC

Both receipts are independently verifiable. No Inntris access required.

What a public receipt proves

Which agent acted — via Ed25519 signature validation
What decision was made — PASS, BLOCK, or ESCALATE
Which policy bound the decision — via policy hash
That the record was anchored — via Base L2 transaction proof

Core capability

Three guarantees. One control plane.

Every agent action passes through three layers: identity verification, policy enforcement, and immutable audit. Together they prove what was allowed to happen, and what actually happened.

Cryptographic identity

Every agent holds an Ed25519 key pair. Every action is signed. No valid signature, no execution.

Policy before execution

Rate limits, spend caps, and action allowlists are evaluated before the action runs — not after the damage is done.

Tamper-evident audit

Every decision is recorded in an append-only log, Merkle-anchored on Base L2. Proves what was allowed to happen, and what actually happened.

Get in touch

Talk to us about your agents.

Whether you're exploring design partner opportunities, have technical questions, or want to know how Inntris fits your agent stack, we respond within 24 hours.

sales@inntris.com

Response time

Within 24 hours

We respond within 24 hours. If you're running agents against code, data, or financial operations — that's our sweet spot.

For design partner discussions, platform reviews, and production agent deployments, contact sales@inntris.com.