Chisle
코딩 에이전트 플러그인, 20개 실시간 작업에서 기본 모델 52% 활용, 간결한 페르소나와 출력 압축 제공
Google Search Console을 Claude와 연동해 URL 기반 32가지 도구를 제공, 읽기 전용 무료
Connect Google Search Console to Claude by URL: 32 tools for quick wins, traffic drops and AI Mode queries across every property. Read-only, free.
개발 도구TypeScript · JavaScriptGitHub @kodotpro온라인운영자 미확인
써 보셨어요? 한 번만 눌러 주세요. 로그인은 필요 없어요.
아직 감지된 업데이트가 없습니다. GitHub 릴리스가 나오면 여기에 보입니다.
Google Search Console MCP server for Claude — 33 SEO analysis tools across every property in your account. Runs locally over stdio or as a remote MCP server with per-user Google sign-in.
README 전체를 불러오는 중입니다.
README 전문 보기 ↗A Google Search Console Remote MCP server that answers SEO questions instead of returning API rows — across every property in your account. Connect it to Claude by URL and sign in with Google. Nothing to install.
https://gsc.k-o.pro/mcp
Paste that into Claude as a custom connector, sign in with your Google account, and ask "which pages lost the most traffic last month, and why?" — you get a diagnosis: a ranking loss, a CTR collapse, or a demand decline. Not a spreadsheet you still have to read.
Guides on gsc.k-o.pro: claude.ai and Claude Desktop · Claude Code · AI Mode queries · self-hosting
Tool reference by group, with parameters and defaults: opportunities · diagnostics · lookups · image search. Every page there is also served as Markdown (add .md to the address, or send Accept: text/markdown); llms.txt indexes them and llms-full.txt is all of them in one file.
A fork with credit due. The 29-tool foundation and the analysis logic behind it are Suganthan Mohanadasan's work, from Suganthan's GSC MCP. This fork adds multi-property access and remote hosting on top. See Credits.
There are several Google Search Console MCP servers, and a growing number of SEO MCP servers generally. Two things make this one different.
It covers your whole account, not one property. Every property-scoped tool takes a site_url argument, and list_properties discovers what your credential can actually see. One install answers questions about all your sites — you never edit a config file and restart to look at a different one. (This is what upstream issue #9 asked for.)
It runs as a remote MCP server, with per-user Google sign-in. Most MCP servers are local-only: one process, on one laptop, with one credential — which is why they cannot be shared, and why the claude.ai and Claude Desktop connector UIs cannot add them at all. This one runs as a hosted service over Streamable HTTP at gsc.k-o.pro: add it as an MCP connector by URL, sign in with your own Google account, and Google's own Search Console permissions decide what you see. Ten people share one deployment and each sees only their own properties — nobody copies a credential anywhere.
It is a full MCP OAuth authorization server: dynamic client registration, PKCE, rotating tokens, and a consent screen of its own. That is what lets the claude.ai and Claude Desktop connector UIs onboard it from nothing but a URL.
Also: it returns SEO analysis rather than rows, every answer states which property it came from, and it exposes Google's type=image surface that most third-party tools skip entirely.
It is built for the people who live in Search Console — in-house SEOs, consultants, and agencies running many client properties from one place.
"What keywords am I almost ranking for?"
"Which pages lost the most traffic in the last month, and why?"
"Do I have any pages cannibalising each other?"
"Compare quick wins across all my properties."
"Which pages get image impressions but no clicks?"
"Is this URL indexed, and if not, why not?"
"Check that 40% CTR figure before I put it in the deck."
Answers carry _meta provenance naming the source, the exact parameters, and the property used — so an answer is never ambiguous about which site it describes. position is labelled as the impression-weighted average it actually is, not a rank-tracker position.
The server is hosted at https://gsc.k-o.pro/mcp. There is nothing to download, no API key to generate, and no config file to edit.
Add to Claude opens claude.ai with the connector pre-filled; you confirm the address before anything connects.
Or by hand: Settings → Connectors → Add custom connector, then paste:
https://gsc.k-o.pro/mcp
claude mcp add --transport http gsc https://gsc.k-o.pro/mcp
Your browser opens and you will see two screens, in this order:
여기까지가 README 앞부분입니다. 나머지는 원문에서 읽을 수 있습니다. README 전문 보기 ↗