# Changelog of Website Technical Audit — SEO, Accessibility & Broken Links (`kikou/audit-technique-site`) Actor

- **URL**: https://apify.com/kikou/audit-technique-site/changelog.md
- **Full Actor documentation**: https://apify.com/kikou/audit-technique-site.md

## Changelog

### 1.2 — 17 August 2026

- The **Website URL** field now comes prefilled with a public crawling sandbox
  instead of a bare `https://`, and the page limit is prefilled at 20 instead
  of 100. You can run the Actor once as-is to see what a report looks like,
  and your first real run on your own site no longer defaults to 100 pages.

### 1.1 — 16 August 2026

**The report is now fully in English, and can be switched to French.**

- Field names and issue codes are now English: `statusCode`, `score`, `title`,
  `imagesWithoutAlt`, `issues[].severity`, code `missing_meta_description`.
  They stay English in every language, so integrations built on them keep
  working when the report language changes.
- New **Report language** setting: English (default) or French. It changes the
  wording of `message` and `label` only.
- Input field names are now English: `startUrl`, `maxPages`, `maxDepth`,
  `includeSubdomains`, `checkExternalLinks`, `concurrency`, `outputLanguage`.
  The previous names are still accepted, so existing saved tasks and API calls
  keep working.
- Key-value store records renamed to `SUMMARY` and `BROKEN_LINKS`.
- New **Accessibility** view in the results table: alt text, declared language,
  link labels, form labels and heading skips on one screen.
- Every issue now carries the measured values in `details`, so you can filter
  and sort on them instead of parsing the sentence.

No change to what the audit detects or to how pages are scored: the same site
audited before and after this release gets the same score and the same issues.

### 1.0 — 16 August 2026

First release.

- 36 checks across SEO, accessibility, mobile, performance, security and
  structured data.
- Broken-link detection, with the page each dead link was found on.
- Duplicate titles and meta descriptions detected across the whole site.
- Summary report: overall score, counts by severity, what to fix first.
- robots.txt is always respected. There is no setting to bypass it.
- Verified against 150 generated sites whose defects were known in advance:
  every injected defect found, no non-existent defect reported.