01 MCP Server · 28 tools

Connect AI to your SEO data

One URL. 27 tools. Works with Claude, ChatGPT, Cursor and any MCP client.

The Model Context Protocol lets your AI assistant call SerpDino directly — check rankings, research keywords, audit competitors, pull traffic data, all from a chat prompt. Same API key as the REST API.

Paid plans only · no credit card to sign up · revoke keys anytime

Get your API key View on GitHub

Claude · SerpDino MCP · 27 tools connected

Show me my projects and how they're ranking this week

list_projects()get_keyword_rankings()

Here are your active projects:

Project	Keywords	Avg position	7d Δ
Acme Corp	142	8.4	+1.2
Blog network	67	12.1	0.0
Client — Pet store	248	5.7	−0.4

Acme Corp moved up 1.2 positions on average this week. Want me to dig into which keywords drove the gain?

Yes, top 5 keywords with the biggest jumpsCompare against our 3 main competitorsGenerate a client report for Acme

Drop this into any MCP client

Endpoint

MCP endpoint

https://serpdino.com/api/mcp

Authentication

Header

Authorization: Bearer sd_your_key_here

Same key as the REST API

Use your SerpDino API key — the same one you'd use for the REST API. Create one in Dashboard → Settings → API Keys.

Three ways to connect

Same API key, different transport. Pick the one that fits your workflow.

Remote MCP

Easiest — paste a URL

Zero install
Always up to date
Best for Claude (web/desktop), ChatGPT

Setup guide ↓

Local MCP

Run as a subprocess

Open-source — fork it
Works offline (against cached data)
Best for Cursor, Claude Code, CI

View on GitHub ↗

REST API

Plain HTTP

Any language, any framework
Full OpenAPI 3.1 spec
Best for scripts, integrations, backends

API docs ↗

Set up in your client

Pick your AI tool — most setups take under a minute.

In the Claude web app, open Customize in the left sidebar.
Go to Connectors and click + Add custom connector.
Set Name to SerpDino and Remote MCP to the URL above.
Open Advanced settings and paste your API key into OAuth Client ID.
Save, then click the ⋯ menu in the top-right and choose Refresh tools list.

27 tools at your fingertips

Every SerpDino capability, exposed as a callable MCP tool. Your AI picks which to call based on what you ask.

Projects

Tool	What it does
`list_projects`	List all your tracked projects
`get_project`	Full details for one project including its keywords
`create_project`	Create a new project with a domain to track
`update_project`	Update project settings, schedule or competitors
`delete_project`	Delete a project and all its tracked data

Folders

Tool	What it does
`list_folders`	List all your project folders
`create_folder`	Create a new folder to organise projects

Keywords

Tool	What it does
`add_keywords`	Add new keywords to track in a project
`remove_keywords`	Stop tracking selected keywords
`refresh_keywords`	Trigger a fresh SERP check — uses balance credits

Rankings

Tool	What it does
`get_keyword_rankings`	Position history across all tracked keywords
`get_serp_results`	Full SERP snapshot for one keyword check
`get_keyword_volumes`	Search volume, CPC and competition data

Research

Tool	What it does
`get_keyword_suggestions`	AI-generated suggestions tailored to a project
`get_keyword_ideas`	Quick keyword ideas from a project's domain

Competitors

Tool	What it does
`get_competitor_rankings`	How a competitor ranks for your tracked keywords
`compare_competitors`	Side-by-side traffic comparison vs competitors

Performance

Tool	What it does
`get_project_pages`	Page-level ranking data with trend sparklines
`get_pagespeed`	Core Web Vitals & PageSpeed Insights for a domain
`get_page_pagespeed`	Per-page Lighthouse metrics for tracked pages
`get_crux_history`	Chrome UX Report (real-user performance) history

Traffic

Tool	What it does
`get_traffic_stats`	SimilarWeb traffic stats for any domain

Notes

Tool	What it does
`list_notes`	Timeline notes & annotations for a project
`add_note`	Add a dated annotation to a project timeline
`delete_note`	Remove a note

Account

Tool	What it does
`get_capacity`	Check account usage, balance and limits
`export_project_report`	Generate a full Markdown SEO report

Things to ask

Click any prompt to copy it, then paste into your AI.

Why use MCP instead of plain REST?

Both work with the same API key. MCP wins when you want your AI to do the work.

Zero-code agent workflows. Skip writing fetch() calls and JSON parsing — your AI picks the right tool and chains them. Build SEO workflows in plain English.
Works in five clients. Claude (web, desktop), ChatGPT, Cursor, Claude Code — one URL, same auth. No per-client SDK to learn.
Same key, no extra billing. MCP calls go through your existing API key. No separate plan, no per-token markup — only your normal SerpDino credit usage applies.
Streaming results. Large datasets (300-row rank reports, full SERP snapshots) stream as they generate. Your AI sees the data as fast as we can fetch it.
Open-source local server. The local MCP server is MIT-licensed. Audit it, fork it, run it inside your VPC — nothing is locked behind our infrastructure.
Built on the REST API you already trust. MCP is a transport — every tool maps to a REST endpoint documented in our OpenAPI spec. Switch between MCP and curl freely.

Frequently asked questions

01What is MCP?

The Model Context Protocol is an open standard from Anthropic that lets AI assistants connect to external tools and data sources. Instead of copy-pasting data into a chat, the AI calls our tools directly and shows you the results.

02Which AI tools is it supported in?

Claude (web, desktop, Code), ChatGPT (Developer mode), Cursor, Windsurf, and any client that implements the MCP spec. We test against Claude and ChatGPT on every release; other clients work via the same standard transport.

03Do I need a paid plan to use the MCP server?

API and MCP access are part of the paid Solo plan. The Free plan gives you the full dashboard — unlimited projects, keywords and update schedules — but no API keys. Subscribe to generate a key and connect your AI tools; you can cancel at any time.

04Does every AI message cost me credits?

No. Read tools (list_projects, get_keyword_rankings, get_pagespeed, etc.) are free — they read from cache and your tracked data. Only refresh_keywords costs balance credits (1 per keyword refreshed). Research tools count against rate limits but not your credit balance.

05Is my data secure when I connect ChatGPT or Claude?

MCP is request/response — the AI only sees data it explicitly fetches during your conversation. We recommend creating a dedicated API key per MCP client (under Dashboard → Settings → API Keys) so you can revoke individual integrations without affecting the others.

06What happens if my AI client is compromised?

Revoke the key in your dashboard — effect is immediate. The client will start receiving 401 Unauthorized on every call. None of your historical data is exposed because MCP doesn't grant persistent access; the client only had what it fetched live.

07Remote MCP vs Local MCP — which should I use?

Remote MCP (a URL) is the easiest path for personal use in Claude web/desktop and ChatGPT. Local MCP (a Node subprocess) is better for IDE workflows like Cursor and Claude Code, or for CI pipelines and custom agents. Both expose the same 27 tools with the same API key.

08Can my AI accidentally delete data?

The tools work exactly as described — delete_project really deletes, remove_keywords really removes. Treat your MCP-connected AI like a power user with full account access. For exploratory testing, create a throwaway project first or use a read-only API key (coming soon).

09How is this different from a Custom GPT or a Claude Project?

Custom GPTs and Claude Projects let you upload static documents. MCP gives the AI live, two-way access to your account — current rankings, fresh SERPs, ability to add/remove keywords. It's the difference between a PDF and a live dashboard.