A documentation audit that runs every night
Every night, HappySupport reads your whole help center and runs 45 checks on it: links, structure, SEO, images, freshness. You get one score, and every finding links to the place where you fix it.
These teams document with HappySupport
Help centers do not break loudly. A linked page disappears, a screenshot ages, a feature changes and its article does not. Each defect is invisible until a reader hits it, and readers rarely tell you.
The usual answer is a quarterly content audit: a spreadsheet, an afternoon per folder, and findings that are stale again by the next release. Checking is exactly the kind of work a machine should do, so your time goes into the fixes.
45 checks across the whole help center
The audit reads everything: every article, folder, redirect, image and setting. External links are actually requested, not guessed at. Screenshots are checked against their age, articles against their traffic and feedback, meta descriptions against what search engines display.
Findings are sorted by severity and by how many readers actually hit the page, so the list starts with what matters. And a clean report tells you what it looked at: articles, links and images, all counted.
What the audit looks at
Every check has a written rationale, a severity and a target. Nothing here is a matter of opinion.
Links
Dead external links, links to unpublished articles, broken anchors, redirect loops. Every URL is really requested.
Structure
Empty folders, folders nested too deep, articles that are really three articles, broken heading hierarchies.
SEO
Missing or duplicate meta descriptions, duplicate titles, thin articles, titles search engines cut off.
Images
Broken images, screenshots older than a year, oversized files that slow the page down, how-tos with no visuals.
Freshness
Popular articles gone stale, drafts nobody published, articles readers vote down, pages nobody reads.
Setup
Custom domain, branding, AI search: the basics that make the help center findable and yours.
A score that tells the truth
Each run produces a score from 0 to 100 with a grade, per area and overall. A category with nothing to measure says so instead of scoring a perfect 100, and no help center gets an A while a critical finding is open.
Findings keep their identity between runs. The audit remembers when each one first appeared, resolves it on its own once it is fixed, and reopens it if it comes back. Snooze what can wait, ignore what you disagree with; both survive every future run.
Fix what the audit finds in one click
The audit surfaces what slips through day to day: dead links and broken anchors, missing or duplicate meta descriptions, duplicate titles, broken heading hierarchies, folders without descriptions or translations, oversized images, screenshots older than a year, articles nobody reads and articles readers vote down.
Much of it you fix directly with AI: it drafts meta descriptions, folder descriptions, clearer titles and translations, repairs headings and compresses images. You review the suggestions and accept them with one click, one by one or all at once.
Your first audit is one address away
Enter the address your customers use today. We import your articles, images and structure as they are, and the first health audit runs right after: every link, image and article checked.
Import from
Zendesk
Intercom
HubSpot
Notion
GitBook
Document360
Help Scout
Freshdesk
Confluence
ReadMe
Mintlify
HelpjuiceRedirects come along, then get checked
The import creates a redirect for every old address. From then on the audit watches them: targets that disappear, loops and chains become findings.
Every link really requested
The first run fetches your external links one by one. Dead ones show up as findings, with the article and the exact link.
Images checked from night one
Broken images, screenshots older than a year and files that slow the page down are flagged, each with its fix.
Finding problems is half the work
The audit tells you what is wrong today. These keep it from going wrong again.
The same audit, applied in public
We run these checks on well-known SaaS help centers and publish the reports.
Your help center, ready before your coffee.
Start for free and publish your first article today.
Book a demo