Website quality monitoring for sites that publish often
For teams that publish often and still check pages by hand. We build a read-only watcher that checks every new and changed page in a real browser, on desktop and mobile, and files ready-to-fix tickets in your tracker.
- Find changesNew and updated pages since the last run
- Check pagesReal browser, desktop and mobile widths
- File ticketsEvidence, severity and a recheck next run
The problem
Manual spot checks stop scaling
Every release can break something a reader notices first: a missing image, a page that scrolls sideways on a phone, a title left untranslated, a share preview without a picture.
Checking by hand catches some of it, on some pages, when someone has time. Problems found late cost more to fix, and the same ones tend to come back.
How it works
How automated website QA works
- 01
Find what changed
Compare your sitemaps and key listing pages with the last run, so attention goes to new and updated pages first.
- 02
Check like a reader
Open each page in a real browser at desktop and mobile widths: loading, images, layout, links, text, titles and share previews.
- 03
File tickets with evidence
Each finding gets a severity, the page address, how to see it, what was expected and a suggested fix.
- 04
Recheck and track
The next run rechecks open tickets and flags any problem that returns after a fix.
What gets checked
What the website checks cover
We agree on the exact list with you, including intentional site behavior that should never be flagged.
| Area | Typical checks |
|---|---|
| Publishing | Pages load, every image appears in order, navigation reaches the right page, and paid or locked content shows as intended. |
| Layout | Agreed desktop and mobile widths, sideways scrolling, and cut-off or stretched content. |
| Content | Typos, placeholder text, missing fields and untranslated text. |
| Search and sharing | Status codes, titles and descriptions, canonical links, language versions, structured data and share images. |
| Whole site | Broken internal links, redirect chains, duplicate titles and page errors, on a rotating schedule. |
Delivered work
Delivered for a digital comics publisher
A quality watcher for a growing comics catalogue
Built with Claude Code for a digital comics publisher with a multilingual catalogue. Each run finds new and changed episodes, checks them in a real browser on desktop and mobile, and files tickets with severity, evidence and a suggested fix. The team reports that checking is easier and faster than reviewing pages by hand, and that a recurring cross-check is now possible.
A digital comics publisher's content team; 2026
Outcome as reported by the publisher's team. Issue counts and time saved are not published.
Published case study
What you receive
A working watcher and a clear handover
Check plan
The page types, language versions, screen widths and reader journeys to cover, and the intentional behaviour to leave alone.
The running watcher
Change detection, browser checks and ticket output, configured for your site and run on the agreed schedule.
Tickets where you work
Findings in the tracker your team already uses, with an export, a severity guide and a short routine for whoever fixes them.
Guardrails
Read-only access, with every fix left to your team
The watcher works with read-only access. It never edits pages, buys anything or unlocks paid content, and a person clears any security challenge.
A page it cannot load or reach is reported as unverified, so a blocked check never looks like a pass.
How engagements work
A written scope and a fixed quote before any build
The founder of AGI Technologies leads every engagement, from the scoping call to the handover.
Every engagement starts with a written scope: the task, what you receive, how the result is judged, the running costs and who maintains it after handover.
The first pilot has a fixed quote, agreed before any work starts. It covers one site, so you can judge the result before anything expands.
Questions
Before a scoping call
Is this uptime monitoring?
It checks what readers see on each page: content, layout, links, search tags and share previews. Uptime alerts are a separate tool, and the watcher can run alongside one.
What do we need to provide?
The page types that change most, the screen widths that matter, access to any signed-in areas you want covered, and the tool where your team tracks issues.
Does it replace our QA team or developers?
It replaces repetitive manual spot checks. Your team still sets priorities and makes every fix; the watcher shows where to look and confirms when a fix holds.
Can it check pages behind a login?
Yes, with an account you provide at the access level you choose. It reads those pages as a user would. It never edits, buys or unlocks anything.
Where does it run?
On a machine you control, on a schedule, or on hosting with its own support plan. Running costs and ownership are part of the written scope. Here is how scheduled Claude Code runs work.
Is it an accessibility or security audit?
It catches common problems readers notice, such as broken layouts and missing text. A full accessibility audit or security review is a separate engagement.
How is the work priced?
After the scoping call you receive a written scope with a fixed quote for the first pilot. Running costs such as hosting and software accounts are listed separately in the same scope.
Scope the next step
Catch publishing mistakes before your readers do
Bring a few pages that change often and the checks your team runs by hand today. On a 30-minute call we scope a first watcher and agree where its tickets should go.