Screpy SEO MCP is a hosted SEO MCP server that connects AI assistants to your Screpy projects through the Model Context Protocol (MCP). It gives your AI client structured access to technical SEO audits, crawl data, Quick Wins, page issues, Core Web Vitals, uptime, and account usage—without copying data between tools or managing a local server.
Important
This is the official integration and documentation repository for the hosted Screpy SEO MCP service. It does not contain the Screpy MCP server source code.
Create a remote HTTP MCP connection in your AI client:
https://mcp.screpy.com
Name the connection Screpy. Do not add an API key, bearer token, or custom authorization header.
Your AI client opens a browser window. Sign in to the correct Screpy account and approve access.
Open a new conversation and ask:
List my Screpy projects.
If the expected projects appear, Screpy SEO MCP is ready.
Screpy runs as a remote HTTP MCP server, so you do not need Node.js, Python, Docker, or a local background process.
| AI client | Connection method | Setup guide |
|---|---|---|
| ChatGPT | Custom app, remote MCP, OAuth | Open guide |
| Claude | Custom remote MCP connector, OAuth | Open guide |
| Cursor | Remote HTTP MCP server, OAuth | Open guide |
| Codex | config.toml and CLI OAuth login |
Open guide |
| VS Code & GitHub Copilot | mcp.json, remote HTTP, OAuth |
Open guide |
| Gemini CLI | CLI configuration, remote HTTP, OAuth | Open guide |
Client menus and remote MCP availability can vary by product version, plan, and workspace policy. The linked Screpy guides are the source of truth for current menu paths.
- Open Settings or Workspace settings in ChatGPT.
- Go to Apps and enable developer mode if ChatGPT asks for it.
- Create a custom app named
Screpy. - Enter
https://mcp.screpy.comand select OAuth when asked how to authenticate. - Let ChatGPT scan the available tools, then sign in to Screpy and approve access.
- Start a new chat and ask:
List my Screpy projects.
View the complete ChatGPT setup guide.
-
Open Claude's Settings or Customize area.
-
Choose Connectors.
-
Add a custom remote MCP connector named
Screpy. -
Enter:
https://mcp.screpy.com -
Complete the Screpy OAuth flow in your browser.
-
Ask Claude:
List my Screpy projects.
View the complete Claude setup guide.
-
Open Cursor's MCP settings.
-
Add a remote HTTP server named
Screpy. -
Enter:
https://mcp.screpy.com -
Do not add a REST API key or custom authorization header.
-
Complete OAuth in your browser, then ask Cursor to list your Screpy projects.
View the complete Cursor setup guide.
Open ~/.codex/config.toml. On Windows, use %USERPROFILE%\.codex\config.toml.
Add:
[mcp_servers.screpy]
url = "https://mcp.screpy.com"Then authenticate from a local terminal:
codex mcp login screpySign in and approve access in the browser. Return to Codex and ask:
Use Screpy to list my projects and show the latest completed crawl for each one.
View the complete Codex setup guide.
Open the MCP server configuration from the Command Palette or Copilot Chat tool settings. Add this remote server to mcp.json:
{
"servers": {
"screpy": {
"type": "http",
"url": "https://mcp.screpy.com"
}
}
}Save the file, start the server from VS Code if prompted, and complete the browser-based OAuth flow.
View the complete VS Code and GitHub Copilot setup guide.
Run:
gemini mcp add --scope user --transport http screpy https://mcp.screpy.comComplete the Screpy OAuth flow when Gemini requests authorization. Then ask:
Use Screpy to list my projects and report the latest completed crawl for each one.
View the complete Gemini CLI setup guide.
Screpy SEO MCP turns authenticated Screpy data into practical, conversational SEO workflows.
| SEO workflow | What you can ask your AI assistant to do |
|---|---|
| Account and usage | Show the current plan, usage, and available features |
| Project discovery | List accessible Screpy projects and identify the correct website |
| Crawl monitoring | Find recent crawls, check status, and summarize completed crawl results |
| Quick Wins | Prioritize the most urgent opportunities from a completed crawl |
| On-page SEO | Group issues by priority and explain the highest-impact fixes |
| Page diagnostics | Investigate status codes, redirects, metadata, and page-level findings |
| Links and images | Find broken links, redirecting links, and image-related issues |
| Core Web Vitals | Summarize performance signals and identify pages that need attention |
| Uptime | Review availability status and recent concerns |
| Project comparison | Compare crawl progress and technical health across projects |
| Crawl actions | Start a crawl only when you explicitly request it |
Available data depends on your Screpy plan, project configuration, completed analyses, feature allowances, and account permissions.
Good prompts name the project or website, identify the relevant crawl when needed, and ask for one clear outcome.
For example.com, find the latest completed crawl and give me an executive
technical SEO summary. Prioritize issues by expected impact and explain the
next three actions.
For my project example.com, find the latest crawl and summarize its most
important Quick Wins.
Give me the on-page audit overview for example.com and group the issues by
priority.
Show pages with HTTP 404 responses from the latest crawl of example.com.
Include the URL, source pages when available, and the recommended fix.
Find redirect chains in the latest crawl of example.com. Prioritize the chains
that affect internal links or important pages.
Show me the crawl data for https://example.com/pricing and summarize the
technical and on-page SEO issues on that page.
Find broken internal links and redirecting internal links in the latest crawl
of example.com. Group them by destination and show the affected source pages.
Review image-related SEO issues for example.com and prioritize the pages that
need attention first.
Which Core Web Vitals need the most work on example.com, and which pages should
I review first?
Give me an uptime summary for example.com and flag any recent availability
concerns.
Create a weekly technical health report for example.com using the latest crawl
data, Quick Wins, Core Web Vitals, and uptime. Separate urgent fixes, trends,
and recommended next steps.
Start a new crawl for example.com.
Starting a crawl is an action. Screpy requires explicit intent and may request confirmation before proceeding.
More examples are available in the official prompt library.
Screpy SEO MCP uses browser-based OAuth. You do not create, paste, or store a Screpy REST API key in your MCP client.
- Access is limited to projects you own or projects shared with your Screpy account.
- Permissions are checked when tools run, not only when the connection is created.
- Screpy does not create, update, delete, or start a crawl unless you explicitly ask.
- High-impact actions require confirmation.
- A Screpy REST API key cannot authenticate an MCP client.
- An MCP OAuth token cannot authenticate the Screpy REST API.
- Do not paste OAuth tokens, browser callbacks, API keys, or other credentials into issues or chat messages.
You can revoke access by removing or disconnecting Screpy from your AI client's MCP, app, or connector settings. See MCP Authentication for the current revocation flow.
| Problem | What to check |
|---|---|
| Screpy does not appear in the client | Confirm that your client or workspace supports remote HTTP MCP servers |
| OAuth does not open | Retry from a local desktop session with a browser and callback path |
| Authentication repeats | Remove the Screpy connection, reconnect, and approve OAuth again |
| Expected projects are missing | Confirm that you used the intended Screpy account and that the projects are owned by or shared with it |
| Tools are missing or unavailable | Refresh or reconnect the server, then verify client and workspace MCP support |
| Data looks incomplete | Confirm that the relevant project has a completed crawl and that the feature is available on your plan |
| A crawl cannot be started | Check project access, account eligibility, current crawl state, and any confirmation request |
| A REST API key does not work | Remove it; Screpy MCP uses OAuth instead |
| Headless login fails | Complete the first OAuth connection from a local computer when SSH, containers, or servers cannot open a browser |
For detailed diagnostics, see Screpy MCP Troubleshooting.
An SEO MCP server exposes structured SEO data and actions to compatible AI assistants through the Model Context Protocol. It lets an assistant use live, authenticated SEO context instead of relying only on text pasted into a conversation.
Screpy SEO MCP is a hosted remote HTTP service at https://mcp.screpy.com. You do not clone this repository to run the server, install dependencies, or maintain a local process.
No. Screpy SEO MCP uses OAuth. REST API keys are for the separate Screpy SEO API and do not authenticate MCP connections.
Screpy provides setup guides for ChatGPT, Claude, Cursor, Codex, VS Code with GitHub Copilot, and Gemini CLI. Any additional client must support remote HTTP MCP servers and browser-based OAuth.
Yes, when your account and project are eligible and you explicitly request the action. Sensitive or high-impact actions may require confirmation.
The connection only sees projects owned by or shared with the authenticated Screpy account. Confirm the account, project sharing, plan eligibility, and feature availability.
Use Screpy SEO MCP when an AI assistant or agent should discover and invoke Screpy tools through OAuth. Use the Screpy SEO API when your own application, backend, dashboard, or automation needs direct JSON API access with a REST API key.
No. This repository is the official documentation and integration entry point for the hosted Screpy SEO MCP service.
- Screpy SEO MCP product page
- Screpy SEO MCP overview
- Setup
- Authentication
- AI client guides
- Example prompts
- Troubleshooting
- Screpy website
- Screpy Security Policy
- Screpy Terms of Service
Use GitHub Issues for documentation and integration problems. For account-specific help, contact Screpy support.
Do not report security vulnerabilities in a public issue. Follow the reporting instructions in the Screpy Security Policy.
This public repository exists to provide a stable GitHub URL for Screpy SEO MCP users, developers, AI client integrations, and MCP directories. The hosted service, authentication flow, and product availability are governed by Screpy's official documentation and terms.
No open-source license is granted by this repository.