Claude Code Plugins

Community-maintained marketplace

Feedback

readthedocs-search-api

@readthedocs/skills
0
0

Query the Read the Docs Search API to find documentation across projects and repositories. Use when searching documentation, finding related docs, finding API documentation, or gathering information about projects on Read the Docs.

Install Skill

1Download skill
2Enable skills in Claude

Open claude.ai/settings/capabilities and find the "Skills" section

3Upload to Claude

Click "Upload skill" and select the downloaded ZIP file

Note: Please verify skill by going through its instructions before using it.

SKILL.md

name readthedocs-search-api
description Query the Read the Docs Search API to find documentation across projects and repositories. Use when searching documentation, finding related docs, finding API documentation, or gathering information about projects on Read the Docs.
metadata [object Object]

Read the Docs Search API

Query the public Read the Docs Search API to find documentation across millions of pages.

Step-by-step instructions

1. Make a search request

Query the API using:

GET https://readthedocs.org/api/v3/search/?q={query}&page={page}

Required parameters:

  • q: Your search query (e.g., "authentication", "API pagination")
  • page: Result page number (default: 1)

Note on scoping: You will generally get better and more consistent results by scoping your query to a specific project. Use fielded search syntax inside q, for example project:{project-slug} your terms. Examples: project:docs build, project:sphinx configuration.

2. Parse the response

The API returns JSON with this structure:

{
  "count": 1234,
  "next": "https://readthedocs.org/api/v3/search/?q=query&page=2",
  "previous": null,
  "results": [
    {
      "id": "project-name:doc-title",
      "project": "project-name",
      "version": "latest",
      "title": "Documentation Title",
      "path": "/en/latest/path/to/page.html",
      "domain": "project-name.readthedocs.io",
      "highlight": "snippet of matching <mark>text</mark>...",
      "blocks": [
        {
          "id": "section-id",
          "title": "Section Title",
          "path": "/path#section-id"
        }
      ]
    }
  ]
}

3. Handle pagination

If there are more results, use the next URL to fetch the next page. Continue until next is null.

Examples

Search for authentication documentation

curl "https://readthedocs.org/api/v3/search/?q=authentication"

Search within a specific project

curl "https://readthedocs.org/api/v3/search/?q=project:docs%20authentication"

Project-scoped quick queries

# Sphinx
curl "https://readthedocs.org/api/v3/search/?q=project:sphinx%20configuration"
curl "https://readthedocs.org/api/v3/search/?q=project:sphinx%20autodoc"

# Requests
curl "https://readthedocs.org/api/v3/search/?q=project:requests%20proxies"
curl "https://readthedocs.org/api/v3/search/?q=project:requests%20authentication"

# Read the Docs
curl "https://readthedocs.org/api/v3/search/?q=project:readthedocs%20build"
curl "https://readthedocs.org/api/v3/search/?q=project:readthedocs%20redirects"

Search with pagination

curl "https://readthedocs.org/api/v3/search/?q=API&page=2"

Parse results with Python

import requests

response = requests.get(
    "https://readthedocs.org/api/v3/search/",
    params={"q": "REST API"}
)

data = response.json()
for result in data['results']:
    print(f"{result['project']}: {result['title']}")
    print(f"  Domain: {result['domain']}")
    print(f"  Path: {result['path']}")

Get all paginated results

def search_all(query):
    all_results = []
    page = 1
    while True:
        response = requests.get(
            "https://readthedocs.org/api/v3/search/",
            params={"q": query, "page": page}
        )
        data = response.json()
        all_results.extend(data['results'])
        if not data['next']:
            break
        page += 1
    return all_results

Common edge cases

No authentication required: The Search API is public and does not require API keys or authentication.

Rate limiting: The API applies reasonable rate limits. If you receive HTTP 429, implement exponential backoff.

Private documentation excluded: Only publicly available documentation is searchable. Private projects are not included.

Search delay: The search index updates with a slight delay. New documentation may not appear immediately.

Empty results: If a query returns no results, try simpler keywords or browse the project directly on Read the Docs.

Project scoping recommended: The global index is large and queries can be broad. For most use cases, include a project:{project-slug} filter in q to scope results to the relevant documentation project.