Numbers for a whole brief, in one call.
whyThe brief lists eighty keywords and no numbers.
asksGet volume, difficulty and intent for the brief's keywords.
Semrush's own MCP server or web app, on a seat of the team's account. What each plan holds back is on this page.
team/tools/semrush/semrush.md plan, seat, projectEstimates and audits, checked against measured clicks.
whyThe brief lists eighty keywords and no numbers.
asksGet volume, difficulty and intent for the brief's keywords.
whyA rival keeps taking the top spot.
asksWhich tracked keywords does our main competitor win?
whyYou want to know if the AI tools mention you.
asksExport our AI visibility from the Semrush app.
With API units, the MCP answers. Without them, the web app does the work.
Positions, landing pages per keyword, and tracked competitors.
Volume, difficulty and intent, up to a hundred per call.
On every plan, with the seat's login.
CSV exports, and AI visibility, which the MCP lacks.
Trends, npm, PyPI, GitHub and Wikipedia, each with its caveat.
Nothing of ours sits in between. No seat? Ask the account holder.
The official MCP server, signed in with OAuth, plus the web app. Without API units, only the final call fails. The browser tools still work.
A page names the plan, seat, project and prompt set.
The seat is yours. Our server stores the plan and project page.
Search Console confirms a drop first.
Starter's cannibalization report is placeholders; Search Console answers instead.
Trends is relative, checked against a control keyword.
What each plan gives, and what it holds back.
| Limit | What it means |
|---|---|
| API units are sold by plan | Without them, only the last call fails. |
| No AI-visibility reports over MCP | Read in the web app. |
| Cannibalization is Pro and up | Starter shows placeholders. |
| Keyword Analytics is not on Semrush One Starter | Every metric reads n/a; tracking and audits still work. |
Name the plan, seat and project. Every AI knows which number is which.