OpenSEO MCP

官方

OpenSEO 提供高品質的 SEO 數據，適用於關鍵字研究、反向連結分析、排名追蹤、網站審查以及分析您的 Google Search Console 資料。其 MCP 伺服器能讓您的 AI 代理更快完成任務，讓您專注於策略規劃。

你可以用 OpenSEO MCP 做什麼？

An SEO MCP server that lets AI assistants pull real keyword, SERP, competitor, and Search Console data mid-conversation.

Keyword research — Expand seed topics into keyword ideas with volume, difficulty, and CPC, then save promising ones to your project.
Live SERP inspection — Check the current Google organic results for a keyword before making recommendations.
Competitor teardown — Pull a domain's overview, ranking keywords, and backlink stats, then summarize where you can compete.
Rank tracking review — Read tracked-keyword positions and latest results from your project's rank trackers.
Search Console signals — Read clicks, impressions, CTR, and position, and inspect URLs for index coverage and crawl issues.
Keyword organization — Group saved keywords by intent and tag them by page or topic cluster for a content plan.

文件

Support our new OSS project, EveryApp. Give it a star→

OpenSEO MCP

An SEO MCP server for AI agents

OpenSEO is an SEO MCP server that connects Claude, Cursor, Codex, or any MCP client to real data, so your agent can research keywords, inspect live SERPs, compare competitor domains, summarize backlink context, save keyword opportunities, review rank-tracking data, and read first-party Search Console signals.

Set up OpenSEO MCP View OpenSEO skills

What is an SEO MCP server?

MCP (Model Context Protocol) is the standard AI clients use to call external tools. An SEO MCP server exposes SEO data (keyword metrics, SERP results, domain and backlink stats) as tools an agent can call mid-conversation. Instead of guessing at search volumes or rankings, your agent queries real data from your OpenSEO project, and can save its findings back so you can review them in the UI. Pair it with keyword research for an agent-driven first pass over any topic.

Agent workflows that work

01 ### First-pass keyword research

Ask the agent to expand seed topics into keyword ideas with volume, difficulty, and CPC, then save the promising ones back to your OpenSEO project for human review.
02 ### Competitor teardown

Point the agent at a competitor domain and have it pull the domain overview, ranking keywords, and backlink stats, then summarize where you can realistically compete.
03 ### Striking-distance sweep from Search Console

Have the agent read your GSC queries, find page-two keywords worth pushing to page one, and check the live SERP for each before recommending changes.
04 ### Keyword clustering and tagging

Let the agent group saved keywords by intent, tag them by page or topic cluster, and hand back a content plan you can act on in the OpenSEO UI.

Available tool groups

Keywords

Research keywords

Get keyword ideas with volume, difficulty, and CPC.
Get SERP results

See live Google organic results for a keyword.
Save keywords

Keep useful ideas organized in your OpenSEO project.
Get rank tracker data

Read tracked-keyword positions and latest results from your project's rank trackers.

Competitive research

Get domain overview

Summarize a domain's organic footprint.
Get domain keywords

Find keywords a domain already ranks for.
Get backlinks overview

Check backlink and referring-domain stats.

Search Console

Get GSC performance

Read clicks, impressions, CTR, and position from the connected property.
Inspect URLs

Check index coverage, crawl, canonical, mobile, and rich-result signals.

AI-Agent SEO strategy library

Practitioner strategies for running SEO through an agent: what to connect, what to automate, where the human stays, and how to make a good run repeatable. Each includes a copy-paste MCP prompt.

Run SEO from your AI assistant: the MCP workflow

Connect one server and the assistant you already use can read Search Console, pull keyword data and check rankings in the same conversation. The first five prompts, and the row that shows why a human still reads the output.
What to automate and what to keep: the dispatcher rule

A scheduled rank check runs without anyone watching. A decision about which keywords go in it does not. Three layers, not two.
Human in the loop content: the brief is the job

Most teams run the loop backwards. Humans write the brief, the model drafts, humans edit, and the two checks that catch the draft that reads like everyone else's.
Skills, memory and the trace: make the good run repeatable

Save the workflow as a skill, give the agent a memory it reads every run, make it write down every step. Plus the two checks that catch the confident wrong answer.

Browse all AI-agent SEO strategies

Google Search Console MCP — no Google Cloud setup

OpenSEO MCP can read Search Console performance and URL inspection data from a connected hosted project. No Google Cloud project or OAuth credentials needed. These tools are read-only and do not use OpenSEO credits.

Explore GSC MCP

Setup lives in Docs

The MCP server URL, Claude setup, Codex setup, and troubleshooting steps are maintained in the docs so this feature page can stay focused on what OpenSEO MCP makes possible.

Open MCP docs

Stay in the loop

Product updates, new features, and the occasional behind-the-scenes.

Email addressSubscribe

OpenSEO

Part of EveryApp

Features

Keyword Research Saved Keywords Rank Tracking Domain Overview Backlinks Site Audit AI Visibility Prompt Explorer All features

AI agents

OpenSEO MCP Google Search Console MCP

Compare

OpenSEO vs DataForSEO Semrush alternative Semrush pricing Ahrefs alternative Ahrefs pricing