A competitor finder AI agents can call is useful for a plain reason: finding rivals is a lookup, then a few follow-up lookups, then a summary. That is the kind of loop a coding agent already runs well. seogeoaeo.ai exposes its Competitor Finder, and the rest of its research tools, over MCP and REST, so Claude Code, Cursor or Codex can run it from the same session where you write code.
What the tool returns
Competitor Finder takes a domain and a country. A full URL is reduced to its domain, so the agent does not need to clean input. It returns the 30 sites sharing the most Google keywords with that domain, with:
- keywords in common and total keywords per site
- the share of each site's keywords that the input domain also ranks for
- estimated monthly visits and average rank
- the input domain's own keyword count and visits
Sites with at least 5% overlap are marked as close competitors. Very large general sites are sorted into their own group. Competition here is measured in search results, not by what a company sells.
Connecting an agent
The setup is short:
- Create an API key at seogeoaeo.ai.
- Add the MCP server at
https://seogeoaeo.ai/api/mcpto your client, or call the REST API underhttps://seogeoaeo.ai/api/v1. - Ask the agent to find competitors for a domain and market.
Tool names, inputs and credit prices are listed at /api/v1/tools and in the site's llms-full.txt, which is a handy file to hand an agent before it plans a run.
What it costs per call
| Tool | Credits | Use in the chain |
|---|---|---|
| Competitor Finder | 38 | Who competes in Google |
| Domain Overview | 50 | Size of each rival: keywords, visits, links |
| Competitor Keywords | 35 | Top 100 keywords for one rival |
| Keyword Gap | 85 | Keywords a rival has in the top 20 that you lack |
Agent runs cost the same as runs in the browser. An agent can set a cap on what a single run may spend, so a loop that goes wrong cannot drain the balance. Each tool shows its price before it runs, and a run that fails before producing a result is not charged. Plans start at $29 a month for 2,000 credits; the pricing page lists the rest.
A sensible chain
A good first prompt for an agent: "Run Competitor Finder on example.com for the US. Take the three closest competitors, run Domain Overview on each, and give me a table." That is 38 plus 150, so 188 credits, and the output is a short ranked list with size estimates.
From there, Competitor Keywords on the strongest rival shows which of its pages bring visits. Keyword Gap shows only what you are missing. Both are named on the Competitor Finder page as the next step.
What the agent will not get
The finder works from Google rankings only. It does not find rivals that sell offline or only run ads, and it does not look at one folder or page. Visits are estimates. A small site with few rankings may get no close competitors, and a lookup that finds the site but no rivals is still charged. There is no rank tracking or ongoing monitoring either; for daily position tracking, a suite such as Semrush, whose SEO plan includes Position Tracking, is the better fit.
For the business side of the picture, the agent can also query the Meta Ad Finder (5 credits) and Competitor Google Ads (15 credits) to see who is paying to reach the same buyers.
Why run it from the editor
Most developers who need this are building a site, a landing page or a docs hub and want to know who already holds the search results they are aiming at. Having the competitor lookup one prompt away means the answer lands next to the code it should shape, instead of in a browser tab you forget about.