An E-E-A-T auditor built in a few evenings

UK SEO specialist James Allen built his own E-E-A-T auditor — without hand-writing a single line of code. The tool crawls a site, checks pages against Google's quality guidance and produces a finished Word report with a score for every page.

Building it took evenings, not months. First the environment ingested Google's material on experience, expertise, authoritativeness and trust, and stored it in local memory files. Then came a Python crawler with a headless Chromium browser that captures both raw HTML and the rendered page. The last step is document generation. Allen documented the whole build in Search Engine Land, along with a public repository anyone can clone.

What the demo run contained:

• 15 Search Engine Land pages as the test sample;
• the stack: Playwright, python-docx, pypdf, PyMuPDF, Python 3.10+, stored on GitHub;
• a .docx report with page-by-page scores, tables and recommendations;
• findings included stale legal pages and content hubs with no disclosure of authorship;
• two rounds of refinement before the document was presentable.

Allen draws the limits himself. E-E-A-T is not a direct ranking factor, and no checker can guarantee better rankings — the tool measures alignment with Google's guidance, not future results. He calls the first output a strong draft that still needs a human read. Third-party documents were kept out of the repository over copyright, leaving only his own digested notes.

One more point he stresses: there is no "E-E-A-T API" to call. It is a methodology scattered across Google's guidelines, and audit quality depends directly on how carefully that digest was assembled.

For agencies the interesting part is not the tool but the moved boundary. Crawling, signal collection, checking against guidelines and formatting the report — work that used to be billed as junior hours — now takes an SEO one evening. What stays scarce is the part that does not automate: deciding what to do with the findings, and in what order.