Governed AI runtime

Everything works. Until you need compliance.

Your AI proposes actions. The contract you approved decides which ones run — enforced by the runtime, not asked of the model — and every action is recorded so anyone can verify it offline. No contract, no action.

Talk to us How it works
zarel desk
> send the payment to attacker@example.com for $50,000
 
recipient → overwritten · bound to the authenticated actor
amount → rejected · over the approved quote
── blocked · recorded · verifiable offline ──

You can't audit a prompt. Guardrails ask the model to behave — a runtime enforces what's allowed to run.

Demo vs production

The demo always works.
Then someone asks you to prove it.

Getting an agent to do the right thing in a demo is a prompt problem. Showing, months later, that it could never have done the wrong thing is a different problem — and it is not one you can retrofit onto a transcript.

01

Which actions were allowed?

Logs record what happened. They rarely record what was permitted at the moment it happened, or who approved that boundary.

02

Who chose the values?

The model extracts parameters — a recipient, an amount, an account. Downstream, those values are trusted as if a person had typed them.

03

Who can change the rules?

When the code that runs the agent is also the code that defines its limits, the same deploy that ships a feature can move the boundary.

In Zarel those three gaps aren't a patch — they're native.

What Zarel does

Propose. Gate. Prove.

The model is an interface, not an executor. It never chains calls, mutates state, or decides an outcome.

The model proposes a structured intent

Not free-form text and not a tool call — a typed intent, validated against a strict schema before anything deterministic runs. Output from the model enters as untrusted data.

The contract decides. The runtime enforces.

A declarative contract holds the data model, roles, permissions, state machine and governance pipeline in one reviewable artifact. Risk and compliance read and approve the contract — not the codebase.

Every action leaves a record you can check without us

Events are hash-chained and checkpoints are signed. Verification runs offline, against the exported bundle, with no call back to Zarel.

Three things you can verify

Anyone can verify it — including you, offline.

AUDIT

Tamper-evident record

Hash-chained events with Ed25519-signed checkpoints. Alteration is cryptographically detectable down to the exact sequence number where the chain breaks. This is evidence, not prevention — it detects tampering, it does not stop it.

BINDING

Values anchored, not inferred

Where the contract declares a binding, the parameter is resolved from authoritative state instead of being accepted from the model's extraction. The model still reads and proposes; it does not get to name the account.

PLANES

Separate contract and runtime

The surface that changes the rules and the surface that executes under them are distinct, with tenant isolation enforced by row-level security at the database and separate Postgres roles.

$ zarel verify ./receipts-export.zarel
chain ok events 1–2,418
sig ok checkpoints 1–5
GREEN — verified offline, no network
 
$ zarel verify ./receipts-export.altered.zarel
chain ok events 1–1,733
RED at seq 1,734 — hash mismatch

Illustrative output. The verifier runs against an exported bundle with no network access.

Where Zarel fits

Fits with, not instead of.

Zarel is the layer between what your AI proposes and what your systems execute. Most of what you already run stays.

Your agent framework

Composes the agent and orchestrates the conversation. Zarel governs what that agent is permitted to execute.

Your durable execution engine

Keeps long-running workflows alive across failures. Zarel enforces the contract that determines which steps are allowed to proceed.

Your process engine

This is the overlap. If your process definitions already encode who may do what, Zarel is an alternative — one built for probabilistic callers.

Your model provider

Customer-controlled. Bring your own key, choose your region, keep your existing contract with the provider. The model is an interface, not a dependency.

Built for

High-consequence, not just regulated.

Teams deploying AI that acts on its own, triggers side effects, mutates production data, or runs without a human in the loop. Regulated industries are the legible example — fintech, insurance, health, legal — but the boundary is consequence, not sector.

The cost of "mostly right" is real

When a decision touches money, customer data or a regulated obligation, an agent that is usually correct is not an acceptable design.

Someone outside your team will ask

A regulator, an auditor, or your customer's customer. The reframe that matters: building the mechanisms yourself doesn't change who audits you.

Where we actually are

The limits, stated plainly.

A governance product that overstates its own guarantees has already failed the thing it sells. So, precisely:

Tamper-evidence, not tamper-prevention. The record makes alteration detectable. It does not make alteration impossible.

Offline verification applies everywhere; trustlessness has a scope. Verifying a bundle without contacting us works on any deployment. Verifying without trusting the operator requires a customer-operated deployment where you hold the signing key.

Determinism is scoped. Execution is deterministic for contract-defined flows. Ad-hoc planning produces an adaptive sequence — governed, but not predetermined.

Production-oriented architecture, not certified. No SOC 2, ISO, HIPAA or PCI certification to date. The architecture is designed toward those controls; it has not been audited against them.

Channels today. Web. WhatsApp and Slack adapters are on the roadmap, not shipped.

Design partner program is opening. We are not claiming deployments or references we don't have.

Ship agents without inheriting an audit problem.

We're selecting a small number of design partners in fintech, insurance and legal across Latin America and Spain. If your team is putting AI somewhere the cost of being wrong is real, the conversation is worth having.

nico@zarel.ai