Skip to main content

Quickstart — Vesanor

Models change without warning. GPT-4o called check_service_health correctly yesterday — today after a silent update, it skips a required tool in the protocol. Your app breaks with no code change. Vesanor catches that.

Start here​

The recommended path is zero-config governance with replay() and Governance Studio.

1. Create your account​

  1. Go to app.vesanor.com/signup
  2. Copy your API key from onboarding. It starts with vsn_
  3. Store it as VESANOR_API_KEY

2. Install the SDK​

npm install vesanor <provider-sdk>

3. Wrap your client​

import OpenAI from "openai";
import { replay } from "vesanor";

const client = new OpenAI();

const session = replay(client, {
  apiKey: process.env.VESANOR_API_KEY,
  agent: "orders-bot",
});

const response = await session.client.chat.completions.create({
  model: "gpt-4o-mini",
  messages: [{ role: "user", content: "Process this order" }],
  tools: myToolDefinitions,
});

That is enough to start the zero-config governance flow.

4. Open Governance Studio​

Once traffic starts flowing, open app.vesanor.com and go to Governance Studio.

There you review:

  • the workspace draft
  • open judgments
  • impact preview
  • approval preview
  • conformance findings after approval

5. Approve​

Approval freezes the current runtime artifact for that workspace and environment.

After approval:

  • development stays non-blocking
  • staging evaluates in advisory mode
  • production enforces protectively and fails closed when governance is unavailable

What happens when something breaks?​

Every failure is classified and fingerprinted so the same failure pattern stays stable across runs. Inspect those failures through Studio, Runs, and References.

Advanced Local​

If you explicitly want local YAML packs, hand-authored contract packs, or recorded local workflows, see the archived docs under docs/legacy/advanced-local/ in the repo.

Next steps​