site-wide seo + aeo audit // powered by typesafe jev
See your whole site,
one dot per page.
dotmap is an SEO checker and audit platform. It crawls your site and scores every page for search and for AI answer engines, then maps what is healthy, what needs work, and what to fix first. Run it here, or from your coding agent.
No account needed.See how it works
Or add it to Claude CodeCursor and VS Code setup
claude mcp add --transport http dotmap https://dotmap.ai/mcp- judgments on every page, from one Jev call
- 12
- technical checks, from title length to JSON-LD
- 18
- to crawl and score 50 pages, measured
- About 3 s
- for those 50 pages, metered while it runs
- $0.005
What dotmap does
Every page, not a sample.
A same-site crawler fetches up to 500 pages and scores each one the moment it lands. Pages it found but did not fetch are listed, so nothing is hidden.
How an audit works 02 · aeoCan an answer engine quote this page?
For each page Jev estimates whether a passage could be lifted as a citation, and flags the rest with a concrete rewrite. It is an estimate, and labelled as one.
What AEO is 03 · compareGoogle and Bing each grade one thing. dotmap grades the whole site.
Neither reads what your pages say. dotmap does, for every page, and shows the Google and Bing numbers beside its own.
dotmap vs Google vs Bing 04 · agentsYour coding agent can run all of it.
dotmap is an MCP server too. Claude Code, Cursor or VS Code can audit a site, read the weak pages, fix them in your repo and audit again.
Connect your agenthow an audit works // crawl → judge → score
Three machines, one audit.
Each figure runs on the same numbers as the audit itself: drag, click or use the keys shown in its card.
crawl // same-site, bounded
A fixed crew of workers, a hard page limit.
The crawler starts at your URL and follows internal links breadth first. A bounded pool fetches HTML pages in parallel and stops at the limit; anything found but not fetched is listed.
- Workers
- 8
- Depth
- 4 links
- Page limit
- up to 500
judge // one jev call per page
One request reads the whole page.
Each page goes to Jev once and comes back with seven scores and five probabilities. Only the input tokens are paid for, so a typical page costs about a hundredth of a cent.
- Judgments
- 12
- Tokens / page
- ≈2,400
- Cost / page
- $0.0001
score // semantic + technical
What a page says outweighs how it is tagged.
Jev's judgments make the semantic score; eighteen checks in code make the technical one. They are blended into one number and a grade from A to F.
- Semantic
- 65%
- Technical
- 35%
- Grade A
- ≥ 85
agents // mcp server
Your coding agent runs the audit, then fixes the page.
dotmap is an MCP server as well as a website. Connect Claude Code, Cursor or VS Code and your agent can audit a site, read the weakest pages with their issues, fix them in your repo, and audit again to check the score moved.
- 01
One command to connect
Add the server to Claude Code, Cursor or VS Code. There is nothing to install and nothing to configure.
- 02
Ten tools, asked for in plain language
Audit a site, list pages by score or issue, compare two runs, turn monitoring on. You ask; the agent picks the tool.
- 03
Sign in once, no API key
Approve access in the browser. A read-only grant can read your reports and cannot start a crawl.
$ claude mcp add --transport http dotmap https://dotmap.ai/mcp
Audit example.com and fix the weakest page.
audit_site25 pages, site score 74
get_pages/features/reports 61, not citable
get_changes2 pages regressed since the last audit
edit app/features/reports/page.tsx+9 -4
Sample session with sample numbers.
alerts // email + slack
An alert the day a page gets worse.
Turn on monitoring for a saved site and dotmap re-audits it daily or weekly, compares the run with the last one, and sends an email or a Slack message when something slips. Your agent can read the same alerts.
- 01
A score drops
By at least the threshold you set, from 1 to 50 points.
- 02
A page breaks
A URL that used to load now returns an error.
- 03
A page stops being citable
An AI answer could quote it before and cannot now.
example.com: score 80 → 74 (-6)
- 2 pages dropped/pricing 91→74, /features/reports 61→52
- 1 page newly broken/old-campaign (404)
- 1 page no longer citable by AI answer engines
report · site history
Sample output
features // after the score
A score is where the work starts.
- Verified fixes
- A new title, description, opening answer, FAQ and JSON-LD for a weak page, re-scored before you see it.
- AI visibility
- Track the questions you care about and see whether an AI answer cites your site, or who it cites instead.
- What changed
- Each audit is compared with the last: pages that improved, regressed or broke, and issues that are new.
- Dot map
- Every page of a saved site as a dot, coloured by score and joined by the links between them.
Pricing // 3 plans
Monthly and yearly pricesFrom the blog // latest guides
All articlesfaq // 6 questions
Questions, answered first.
Direct answers up top, the way this tool asks every page to do it.
All questionsdotmap is a site-wide SEO and AEO audit. It crawls pages up to the selected limit, has TypeSafe's Jev model judge each fetched page on twelve dimensions in a single call, runs eighteen technical checks, and produces a shareable report with a score per page and for the whole site. Signed-in users can run a deeper sitemap-assisted crawl and explore a private dot map. Coverage limits are shown in the report.
ready
Your site, scored, in the time it takes to read this.
Start without an account. Sign in later for saved history, monitoring and your private dot map.