Until now the content audit lived only in the web app. The technical audit was on the API and the MCP server already, but the scores and findings for your copy were not. If you wanted Claude Code to improve a page, you had to paste the findings in by hand.
Now your agent can read it all directly.
What is new
4 MCP tools, in aeo-copilot-mcp 2.1.0:
get_content_audit lists every audited page with its quality, substance, trust and copy scores, the last change, and how many prompts point to it.
get_content_audit_page returns everything about one page: the full report (each tell with its fix and the lines it found), the score history, the prompts that point to the page, and its open actions.
set_content_audit_pages picks which pages are audited, up to 20.
run_content_audit rescores the pages after you edit them.
The same data is on the public API:
Why it matters
The pages worth fixing first are the ones your prompts point to. Your agent can now find the weakest of those, read why it scored low, rewrite the copy on your site, and check the new score, all in one session. You review the edit instead of writing the brief.
Access follows your role on the brand: viewers can read, editors can change pages and rescore. These tools need the new version of AEO Copilot on the brand.
Run npx aeo-copilot-mcp@latest to get 2.1.0. Setup steps are in how to connect AEO Copilot to Claude.