Connect meev

Bring your data into the agent you already use

Connect meev to your AI tools with a live connection or an installable skill. Pick your tool below, connect once, and ask about your rankings, AI visibility, and competitors in plain language. Both are read-only, so they can look but never change anything.

Live connector

1

Copy the meev address

You'll paste it into Claude in the next step.

https://meev.ai/api/mcp
2

Open Settings, then Connectors

Add a custom connector, name it meev, and paste the address.

3

Connect and sign in

Click Connect, sign in to meev, and approve. Now ask Claude about your rankings.

Want the full walkthrough, example prompts, and limits? See the MCP page
Preview the meev-data skill
---
name: meev-data
description: Answer questions about the user's own search and AI-search performance by calling meev's read-only REST API (meev.ai) — rankings, AI visibility and citations, competitors, backlinks, content, topical authority, and keyword gaps for the domains the user owns. Use when the user asks where they rank, whether AI answers mention or cite them, who is outranking them, which pages lost traffic, or what to write next.
---

# meev data

Answer the user's questions about their search + AI-visibility performance by
calling meev's read-only REST API. Every result is scoped to the domains the
user owns. The API can only read: it never edits, publishes, or spends, so never
tell the user you changed anything on their behalf.

## Setup (once)

1. The user mints an API key at
   https://meev.ai/dashboard/settings/api-keys — it's shown once, so they copy
   it immediately. Store it as `MEEV_API_KEY` in the user's secrets manager
   (`.env.local` for local dev, the platform's env store for deployed apps).
   Never commit it or print it back in full.
2. Base URL: `https://meev.ai/api/public/v1`
3. Send the key on every request as an `x-api-key` header (or
   `Authorization: Bearer <key>`).

## How responses work

- Success: `{ "success": true, "data": ... }`.
- Failure: `{ "success": false, "error": { "code": "...", "message": "..." } }`
  with the matching HTTP status (400 bad request, 401 missing/invalid key,
  403 the key doesn't own that domain, 404 unknown resource, 429 rate limited).
- Always start by listing domains to get a `domainId`, then query the
  per-domain resources with it.
- The full, always-current machine-readable reference is
  https://meev.ai/api/public/v1/openapi.json — fetch it to discover every
  endpoint and its exact response shape. Treat it as authoritative over the
  list below.

## Core endpoints

Account & domains:
```
GET /domains                              the websites the user manages (get the id)
GET /domains/{id}                         business, voice, and positioning profile
GET /account                              plan, limits, and usage
```

Search & SEO:
```
GET /domains/{id}/rankings                organic ranking posture + top keywords
GET /domains/{id}/search-performance      clicks, impressions, and queries
GET /domains/{id}/keyword-gaps            keywords competitors rank for that you don't
GET /domains/{id}/index-coverage          which pages are indexed vs not
```

AI visibility:
```
GET /domains/{id}/ai-visibility           mentions + citations across AI search
GET /domains/{id}/ai-visibility/citations which sources AI answers cite
```

Competitors, backlinks & content:
```
GET /domains/{id}/competitors             who wins the queries you want
GET /domains/{id}/backlinks               referring domains + authority
GET /domains/{id}/content                 themes, cadence, and coverage
GET /domains/{id}/topical-authority       your topical map + gaps
```

## Example request

```
curl https://meev.ai/api/public/v1/domains/{id}/rankings \
  -H "x-api-key: $MEEV_API_KEY"
```

## Rules

- Read-only. Never claim you edited, published, or spent anything.
- Only the user's own domains resolve; a 403 means the key doesn't own that
  domain, not that you should retry.
- If a resource comes back empty, say the data isn't captured yet. Do NOT
  invent numbers.
- Answer from the JSON: cite the concrete figures (rank, delta, clicks, %),
  don't hand-wave.
- For anything not covered above, read the OpenAPI spec before guessing a path.

Two ways in, one set of data

Both read the same growth data. Pick the one that matches how you work.

MCP

Live connector

One live connection for every assistant. Paste an address into Claude or ChatGPT, or add it with one command in Claude Code and Cursor, then sign in.

Skill

Agent skill

Install a skill file so your agent pulls your data through the API, using a key you create.

What your agent can see

The same growth data meev tracks for you, available to the agent you already work in.

Search rankings & performance

Where you rank, what's moving, and the queries driving clicks and impressions.

AI visibility

Whether AI answers mention and cite you across every major AI search surface: ChatGPT, Claude, Gemini, Perplexity, Grok, Google AI Overviews, and AI Mode.

Competitors & backlinks

Who's winning the queries you want, and where their authority comes from.

Content & topical authority

Your themes, pillar coverage, publishing cadence, and where the gaps are.

Keyword gaps & next steps

The keywords competitors rank for that you don't, and what to do about them.

Read-only by design. Both methods can look, but neither changes, publishes, or spends on your behalf. Each request is scoped to the domains you own, and you can disconnect anytime.

Ask it anything

No dashboards to dig through. Just ask, the way you already talk to your agent.

  • Which of my keywords slipped this week?
  • Where do competitors outrank me?
  • How is my AI visibility trending?
  • What should I write next?
  • Which pages lost traffic?

Building against your data with code instead? Read the developer API docs.

Frequently asked

What's the difference between MCP and the skill?
MCP is a live connection. In apps with a connector setting, like Claude or ChatGPT, you paste an address and sign in; in coding agents like Claude Code and Cursor, you add it with one command. The skill is a file you install so a coding agent pulls your data through meev's API with a key you create. Both are read-only and read the same data.
Which one should I use?
Use MCP in most cases: you sign in through your browser and there's no key to manage. Install the skill when you'd rather your agent call the API with your own key.
Which AI assistants can I connect?
Claude Desktop, Claude web, and Claude Code, plus Cursor, ChatGPT, and any client that supports the Model Context Protocol.
Do I need an API key?
Only for the skill. MCP signs you in through your browser, so there's no key to manage. The skill uses an API key you mint in your dashboard.
Can it change or publish anything?
No. Both methods are read-only. Your agent can read your data to answer questions, but it can't edit, publish, or spend on your behalf.
Which plans include it?
Every paid plan. Higher plans allow more requests and more connected agents.
How do I disconnect an agent?
Anytime. Remove a connected agent from Settings then Connected agents, or revoke an API key from Settings then API keys. Either one cuts access immediately.

Ready to ask your agent about your own growth?

Get started