Run an audit

Tell us your brand and the questions you want to win. We ask the engine, check the sources, and hand you a score plus the files to fix it.

Brand name *

Website URL *

Keywords * (1–5)

Competitors (optional, 0–3)

Brand aliases (optional, 0–5)

Product names your brand is known by. If an AI answer names them, it counts as mentioning you.

Engine *

AI modelAny OpenAI-compatible key · fast · cheap · a first pass Web searchPerplexity · live search · real citations · verifiable

API key * Any OpenAI-compatible key works — from OpenAI, OpenRouter, or any relay. Paste it, run. Stored only in this browser.

No key yet? Get discounted API credits from a recommended provider:

Your AI visibility score

A single number: how often AI answers mention your brand when people ask about your keywords.

Run an audit to see where you stand.

Score—

Cited—

Engines—

Verdict—

Engine verdicts

What the engine actually said, keyword by keyword — with the sources it named.

Who wins the answer

When AI recommends a category, does it name you or a competitor?

Brand	Cited	Share	Verdict

What AI says about you

A plain-language summary of how you come across in the answers.

—

The fix pack

Ready-to-ship files that help AI engines find and cite you. No dashboard, no subscription — just files.

llms.txt
AI site map

robots.txt
Crawler check

Schema
FAQ JSON-LD

Capsules
Answer drafts

llms.txt — deploy at yourdomain/llms.txt

Run an audit and your llms.txt will appear here.

robots.txt — AI crawler check

Paste your robots.txt (or leave it blank) and we will check if AI crawlers are blocked — then give you a fixed version.

FAQPage JSON-LD — schema markup

Schema markup that helps AI engines quote your pages. Built from your keywords.

Run an audit and your schema markup will appear here.

Answer capsules — content drafts for missed keywords

Short, ready-to-publish answer paragraphs for the keywords you missed — the format AI engines like to quote.

Run an audit and your answer capsules will appear here.

Agent-readiness check

AI coding agents like Claude Code, Cursor and Codex read specific files to learn about your product. Here is how findable you are to them.

Fix plan

Your to-do list: what to change, in order of impact.

The HTML report is clean and client-ready — open it, print to PDF, send it. JSON has the full raw data for your own tools. Sources the engine listed are checked live in your browser; dead or unreachable ones are flagged so you are never misled.

How it works

[*]

You bring the questions. The engine answers. We check the receipts and hand you the files.

3 samples per keyword

the same question asked three ways, so one lucky phrasing can't fake a win.

Citation checks

every source an answer cites is fetched live. Dead pages and fake mentions get flagged.

Position tiers

lead, top 3, listed, or absent. Where you appear changes the score.

Calibration control

a nonsense query is included; if the engine "cites" you for it, confidence is capped.

Score, not vibes

cited answers ÷ total answers, turned into a 0–100 score with a verdict.

Fix files, not advice

llms.txt, robots.txt rules, FAQ schema and answer drafts, ready to ship.

Private by design

your key never leaves the browser. No account, no server-side logs.

Client-ready export

a clean white-label HTML report, print-ready.

FAQ

Get better AI visibility, first

New checks and fix files, delivered to your inbox before anyone else sees them.