Analytics for AI search

Know how AI engines answer for your brand.

Continuous visibility intelligence across ChatGPT, Claude and Perplexity — whether you're mentioned, who's cited instead, and which sources the models trust.

Start tracking See how it works

Self-hostable · own your data · no scraping, official engine APIs only

Tracking visibility across the engines that shape answers

How it works

From a domain to a visibility report in three steps.

No tags to install. You describe what you want to rank for; we probe the engines on a schedule and turn the answers into metrics.

Connect domain & prompts

Add your domain, brand names and competitors, then a set of prompts your buyers actually ask. Manage them in the dashboard or via the API.

We probe the engines

On your schedule, each prompt is sent to every engine through official APIs. We capture the full answer and its citations — no scraping.

Read visibility & act

See visibility rate, share-of-voice vs competitors and which sources the models cite — with concrete recommendations and ready-to-use prompts.

The dashboard

Everything in one quiet view.

Why it's built this way

Own the data. Automate the analysis.

Own your data, self-host it

Open-core, in the Plausible tradition. Run it on your own infrastructure or use the cloud — either way the probe history is yours to keep, export and query.

Postgres + ClickHouse · full export

API · CLI · MCP first

Every metric is reachable from a query-style API, a CLI for your pipelines, and an MCP server so your own agents can read visibility directly. The dashboard sits on top, not in the way.

geo visibility · geo gaps · geo recommend

A dashboard that stays simple

One page, deliberately. Visibility over time, share-of-voice, the domains each engine trusts and where you're missing — without a wall of charts you'll never open.

Single-page, fast, no setup

Recommendations, not just charts

For every prompt where a competitor is cited and you aren't, you get the gap, the sources the model leaned on, and an optimization prompt you can act on today.

Gap analysis · ready-to-use prompts
Developer-first

If you'd rather not open the dashboard.

Every metric is reachable three ways — a query API, a CLI for your pipelines, and an MCP server so your own agents can read visibility directly.

  • API — the same query shape as the dashboard. Bearer auth, scoped keys.
  • CLI — drop a visibility check into CI or a nightly cron.
  • MCP — let Claude or any agent pull share-of-voice on demand.
# Same query shape as the dashboard
curl https://api.usementio.eu/v2/query \
  -H "Authorization: Bearer $MENTIO_KEY" \
  -d '{
    "site_id": "acme.com",
    "metrics": ["visibility_rate", "share_of_voice"],
    "dimensions": ["geo:engine"],
    "date_range": "30d"
  }'
# Pipe visibility into CI or a cron job
$ geo visibility --site acme.com --last 30d

ENGINE       VISIBILITY  CITATION  SHARE
all          61%         38%       34%
chatgpt      67%         41%       38%
claude       58%         36%       31%
perplexity   54%         39%       30%
// Add the MCP server to your agent
{
  "mcpServers": {
    "mentio": {
      "command": "npx",
      "args": ["-y", "@mentio/mcp"],
      "env": { "MENTIO_KEY": "sk-..." }
    }
  }
}
// → "Where am I losing share of voice?"
Why it matters

Stop guessing what the models say about you.

Checking by hand

  • Prompt an engine now and then, screenshot the answer
  • No history — you can't see a trend or catch a regression
  • Guess which competitors get named in your place
  • No idea which sources the model actually trusts

With

  • Continuous probing on your schedule, across every engine
  • Full history you own — query any window, any dimension
  • Share-of-voice against each competitor, over time
  • The exact domains each engine cites — your optimization map
Pricing

Simple plans. Self-host free, or let us run it.

Self-host the full engine free under AGPL, or pick a managed cloud plan — reports included on every plan. See full pricing →

Cloud

Starter

For one brand, getting your first signal.

$19
per month
  • 25 prompts tracked
  • ChatGPT + Perplexity
  • Weekly tracking · 1 site
Get Starter
Cloud · popular

Pro

For brands serious about AI visibility.

$69
per month
  • 50 prompts tracked
  • ChatGPT + Perplexity + Google
  • 3 sites · competitor + reports
Get Pro
Cloud

Premium

All engines, more sites, the full picture.

$199
per month
  • 100 prompts tracked
  • All 4 engines, incl. Claude
  • 10 sites · priority scheduling
Get Premium

Also free self-host (AGPL), pay-as-you-go and Enterprise — compare all plans →

FAQ

Questions worth asking.

Do you scrape ChatGPT or Perplexity?

No. We only call the engines' official APIs — the ones that return answers with citations. There's no scraping of the consumer apps, which keeps the data clean and the terms of service intact.

Can I really self-host it?

Yes. The core is open and runs on your own infrastructure (Postgres + ClickHouse), in the Plausible tradition. The cloud edition is for teams who'd rather not operate it — but the probe history is always yours to export and query.

Which engines are supported?

ChatGPT, Claude and Perplexity at launch — the engines with official APIs that expose citations. Google AI Overviews is next, via licensed providers. We add engines where we can do it cleanly, not by scraping.

How are the prompts chosen?

You define them — the questions your buyers actually ask. You can manage them in the dashboard or the API. Automatic prompt-set suggestions, anchored to real search demand, are on the roadmap.

Is AI search even worth tracking yet?

Honestly, referral traffic from AI answers is still a small slice of the web today. We don't sell you a traffic miracle. What we sell is visibility intelligence: the answers are being written now, and knowing where you stand is cheap insurance before it matters.

How is this different from an SEO tool?

SEO tools rank pages in search results. We measure whether the answer itself mentions and cites you — across multiple engines — and which sources those models lean on. Different surface, different metric.

Start measuring what AI search already says about you.

AI referral traffic is still a sliver of the web today — but the answers are being written now. Know where you stand before it matters.

Early access · no spam · self-host anytime without us.

Read the docs