Great content on a broken site loses
Technical SEO is the plumbing: invisible when it works, catastrophic when it doesn't. A robots.txt that accidentally blocks all crawlers — a single character's difference from a normal one — can remove a site from search entirely, while its owner keeps publishing into the void and wondering why nothing ranks.
SEOtopilot generates content for your site, so it checks the site too. Publishing thirty fact-checked articles onto a domain that crawlers can't read would be malpractice we'd be participating in.
What the audit checks
Working from the outside like a crawler would, the audit covers:
- HTTPS — served securely, full stop. A critical finding if not.
- robots.txt — present, parseable, and not blocking everything. The accidental global Disallow is rarer than folklore suggests but genuinely fatal, so it's checked first and flagged critical.
- sitemap.xml — present, valid XML, actually containing URLs, and referenced from robots.txt. Malformed and empty sitemaps are distinct findings, because their fixes differ.
- Meta fundamentals — title present and within length, meta description present and within range, canonical link, mobile viewport. The unglamorous basics that broken themes quietly drop.
- Structured data — JSON-LD present and syntactically valid, with a declared type. Invalid structured data is worse than none: it signals carelessness to the one audience that parses it strictly.
- llms.txt — the emerging convention for guiding AI crawlers. Forward-looking, and noted rather than penalised heavily — but if you care about AI visibility, it's table stakes arriving early.
A score you can act on, not a data dump
Audit tools love volume — six hundred rows of findings, exported to a spreadsheet, actioned never. SEOtopilot returns a health score from 100 with findings in three severities:
- Critical — actively costing you search presence. Fix today.
- Warning — meaningful, fix this week.
- Notice — worth knowing, batch when convenient.
Findings are sorted, weighted, and counted; the score is deterministic, so fixing issues visibly recovers it. The judgement of what to do stays with you — the same division of labour as the review workflow, applied to infrastructure.
Audits and content, one loop
The audit isn't a separate product bolted on for the feature list. It runs on a schedule, and its findings sit alongside your content calendar — because "fix the malformed sitemap" sometimes outranks "publish article #31" as the best use of a week, and a tool that only ever recommends more content can't tell you that.