[Bug]: Error checking public website during Context7 submission

[freddydcg]

MCP Client

Other (specify in description)

Context7 MCP Version

N/A — remote Streamable HTTP MCP endpoint; no local @upstash/context7-mcp package installed

Bug Description

MCP Client: Other (ChatGPT Desktop / Codex Desktop)

Submitting the public Pilot documentation website to Context7 fails with “Error checking website”.

The Context7 API key is valid: library searches succeed. However, submitting either the website root or its static documentation index does not return a library identifier.

This appears to be a website validation/crawling issue rather than an authentication issue.

Steps to Reproduce

1. Configure Context7 as a remote Streamable HTTP MCP server at:
   https://mcp.context7.com/mcp

2. Authenticate with a valid Context7 API key stored securely outside project files.

3. Confirm authentication works with:
   GET /api/v2/libs/search

4. Submit the public website through:
   POST /api/v2/add/website

   {
   "websiteUrl": "https://doc.pilot-gps.com/",
   "websiteBaseUrl": "https://doc.pilot-gps.com/"
   }

5. Retry with the static documentation index:

   {
   "websiteUrl": "https://doc.pilot-gps.com/contents.html",
   "websiteBaseUrl": "https://doc.pilot-gps.com/"
   }

Expected Behavior

Context7 should validate and submit the public documentation website successfully, returning a library identifier for later documentation queries.

Actual Behavior

Both website-submission requests take approximately 30 seconds and fail. No library identifier is returned.

The API key is valid because library searches succeed.

Both Pilot URLs are public and return HTTP 200. The static index contains the Pilot user-guide table of contents and links to its documentation pages.

Error Messages / Logs

{
  "error": "other",
  "message": "Error checking website"
}

Transport Method

http

Node.js Version

N/A for this issue — remote HTTP MCP. Node.js v25.2.1 is installed but not used.

Operating System

Windows 11 Home 10.0.26200

Configuration

[mcp_servers.context7]
url = "https://mcp.context7.com/mcp"
http_headers_helper = "<secure local helper omitted>"
startup_timeout_sec = 30
tool_timeout_sec = 60

Additional Context

• Transport is remote Streamable HTTP, not stdio.
• No local @upstash/context7-mcp package is installed or used.
• The website does not expose llms.txt, sitemap.xml, or robots.txt at conventional paths.
• Please advise whether the domain is blocked by the website validator, requires crawler allowlisting, or needs a specific entry URL or metadata file.

[modelpath-dev]

I will take this issue. Please assign it to me.

The problem seems to be with the website validation or crawling process. The error message "Error checking website" suggests this. I will first check the code handling the POST request to /api/v2/add/website. I will look for any conditions or exceptions that might cause this error. A small fix might involve adjusting the validation logic or improving error handling to provide more specific feedback. I'll also verify if the absence of llms.txt, sitemap.xml, or robots.txt affects the process.