documentation

Vorza documentation

Vorza measures whether AI agents can actually use your product — and refuses to blur evidence with measurement. The pre-check is a static scan: 42 checks against the surface your site presents to an agent, scored in seconds, every failing check shipping a concrete fix. The audit is the measurement: real Claude Code and Cursor sessions run against your MCP server, SDK, or CLI, repeatedly, in isolated sandboxes — reported as per-scenario success rates with 95% Wilson confidence intervals and downloadable evidence behind every run.

After the audit, the loop closes: the Fix Loop hands findings to your own coding agent over MCP and verifies fixes by re-running the exact failing scenarios, and monitoring re-runs your plan on a schedule so a client release can't quietly break what passed.

Start here

Quickstart

The whole loop in five steps: scan, audit, read results, fix, verify.

Pre-check scan

How the scan works

Four layers, weighted scoring, grades — and what a static scan can and cannot claim.

All checks

Every check the scanner runs, by layer, with its weight and generic fix.

Audits

How audits work

Real Claude Code and Cursor sessions against your product — the full lifecycle.

Plans & assertions

How an evaluation plan is generated, what a scenario contains, and the assertion catalog.

Credentials

Sealed-box encryption in your browser, worker-only decryption, scrubbing, purge.

Results & findings

The verdict, the matrix, AI-drafted findings with evidence, and your free re-run.

Scenarios

Framework starting environments, the scenario builder, cycles, and the verified framework-by-agent matrix.

Preference

Do coding agents choose you? Markets, brand detection rules, share cycles, matchups, and the ranking board.

AEO

How retrievable and citable is your brand to agents? The deterministic check catalog, the honest crawler, source attribution, gaps, and the closure contract.

Experience

Can agents use you? Evals per integration path, judged runs, the onboarding funnel, time to first success, and regression watch.

Fix & monitor

Fix loop

Findings become durable threads: your agent pulls specs over MCP, fixes in your repo, and the re-run's wire evidence flips the verdict.

Agent readiness

The reference library: MCP server design, in-band error remediation, agent-readable docs, one-shot quickstarts, and more.

Monitoring

Your audit's plan, re-run on a schedule; email only when something regresses.

Publish

Publishing

Opt a subject into a public scorecard, comparison page, framework page, or verified-fix changelog, with the live badge and the evidence standard behind every number.

Reference

MCP servers

The public server (scan, scores, plans) and the customer-scoped audit server.

HTTP API

The endpoints behind the site — openapi.json is the authority; this is the tour.

For agents

llms.txt, agents.md, skill.md, mcp.json — everything an agent can read without asking.

reading this as an agent?

Everything here has a machine-readable twin: /llms.txt, /agents.md, /api/openapi.json, and the MCP advertisement at /.well-known/mcp.json. See Machine-readable surfaces.

Loading…