The record agents read, in your hand.
In one sentence
The CrawlCheck app is the companion to crawlcheck.io, live on the web at crawlcheck.io/app/run with store listings in preparation, that reads the same dated record the site keeps and the signed Resolve answer an agent gets for each domain, follows the site's design tokens and section list live from /tokens.css and /app.json, and adds a field kit of instruments that only a phone can run — the third vantage after the datacentre reading and the browser extension, each instrument sending a receipt back to the record and none of them computing a score.
What an agent is told about your site, on the home screen. A receipt the moment that answer moves. And a field kit of instruments that only work in a hand, on a street, or on a network the datacentre will never stand on — each one adding evidence to the same signed record agents read.
Open the app →The other client: the browser extension →
6
instruments at launch
35
sections, the same as the site
0
scores computed on the phone
A native companion to the record
Nothing about a score is computed on the phone: the app reads the record the site keeps, and adds to it.
Same record, same vocabulary
It scans a domain the way the site does and shows the audit on the same report vocabulary — Reach, Read, Quote, the same 35 sections, the same grade — beside the dated record and the signed Resolve answer an agent gets for that domain: a decision for read, cite, connect and transact, never a composite number.
Instruments a server lacks
A phone has a camera, a location, a cellular network and a share sheet, and each of those is an instrument here. Every instrument sends what it measured back as a receipt, so the record gets sharper every time one is used — the third vantage after the datacentre reading and the browser extension, and the only one that can stand at the address.
Follows the site live
Everything the app paints — colour, type, the section list, the plans — it takes from two files this site already serves, so it follows the site the moment the site changes and hard-codes none of it.
What a phone can do that a server cannot
Six instruments ship at launch. Fourteen more are in the design and appear in the app as roadmap — visible, explained, with a notify-me — rather than as promises on this page.
Citation rounds
Ask the engines from your own phone. Share any answer back; the app files whether you were cited, who was cited instead, and the date.
Network replay
Re-run the identity check from your own network. A datacentre reading and a cellular reading of the same page, side by side, on the record.
Field proof
Stand at the address. The camera checks the sign and the hours against the schema; the evidence is hashed into the record.
Deploy handshake
A fix ticket that closes itself. The developer scans the QR when they ship; the app re-measures within the minute and tells you both.
Receipts inbox
Every dated change, as a push. Snooze it, assign it, or mark it known; each one deep-links to the finding with the byte diff open.
Scan from anywhere
A URL in the share sheet becomes a scan; a screenshot of an answer becomes a citation receipt; a log becomes a verifier run.
What the site serves for the app
77
design tokens at /tokens.css, light and dark
35
sections in /app.json
4
plans in /app.json
8
app API routes live
The app follows /tokens.css (a stable alias of the site’s own design tokens, 5-minute cache, ETag) and /app.json (sections, stage questions, plans, limits, checkout), both served from this site’s Worker with CORS. Its backend is live too: sign in with the licence key from your receipt, read your portfolio and each domain’s record, register for push, post field-instrument measurements, and send every decision the phone makes as a receipt to /api/app/receipts. None of those routes can move a score; they read the record and add receipts to it.
What it measures, and what it never does
Measured
About 30 read-only requests to your public pages and machine files, sent under the scanner’s own name, plus one homepage fetch under each named crawler user-agent. Beyond that, only what an instrument you enabled asks for, only while it runs.
Never
No writes to your site or listings. No scraping of people. No selling your record. No location outside an instrument that needs it, and none kept at rest — a location is written into the receipt it belongs to, then discarded.
Where it stands
Free with any plan. A plan today is not waiting on the stores: the record keeps measuring twice a day, and the app opens onto whatever that record holds.
- 2026-09-05
Design frozen
28 screens, the field kit, the live-token contract.
- 2026-09-17
Android package built
The store listings are in preparation; this page will say so until they are live.
- Live today
crawlcheck.io/app/run
Open it on your phone and add it to the home screen. It installs like an app, works offline for the shell, and every scan is a real request to this site.
Open the app on your phone
Scan any domain free; sign in with the licence key from your receipt to open your portfolio and the record.
The browser extension →The live demo report →The API →