docs-manage

Docs Manage

Index, refresh, and remove library documentation in the local Grounded Docs store. These commands modify the index and produce plain-text status messages on stdout.

When to use

• A library is not yet indexed and you need its docs available for search.
• Documentation may be stale and you want to pull in updated pages.
• You want to remove a library or version from the index to free space.
• The store file is still large after removals and you want to compact it.

Commands

scrape

Download and index documentation from a URL or local directory.

npx @arabold/docs-mcp-server@latest scrape <library> <url> [options]

Flag	Alias	Default	Description
--version <ver>	-v		Library version label
--max-pages <n>	-p	config default	Maximum pages to scrape
--max-depth <n>	-d	config default	Maximum navigation depth
--max-concurrency <n>	-c	config default	Concurrent page requests
--ignore-errors		true	Continue on individual page errors
--scope subpages|hostname|domain		subpages	Crawling boundary
--follow-redirects		true	Follow HTTP redirects
--no-follow-redirects			Disable following redirects
--scrape-mode auto|fetch|playwright		auto	HTML processing strategy
--include-pattern <glob>			URL include pattern (repeatable)
--exclude-pattern <glob>			URL exclude pattern (repeatable, takes precedence)
--header "Name: Value"			Custom HTTP header (repeatable)
--embedding-model <model>			Embedding model configuration
--server-url <url>			Remote pipeline worker URL
--clean		true	Clear existing documents before scraping
--replace		false	Rebuild the library version if it already exists
--quiet			Suppress non-error diagnostics
--verbose			Enable debug logging

Examples:

# Scrape React docs, version-tagged
npx @arabold/docs-mcp-server@latest scrape react https://react.dev/reference/react --version 19.0.0

# Scrape local files
npx @arabold/docs-mcp-server@latest scrape mylib file:///Users/me/docs/my-library

# Scrape with depth and page limits
npx @arabold/docs-mcp-server@latest scrape nextjs https://nextjs.org/docs --max-pages 200 --max-depth 3

# Scrape with custom headers (e.g. authentication)
npx @arabold/docs-mcp-server@latest scrape internal-api https://docs.internal.com \
  --header "Authorization: Bearer tok_xxx"

# Exclude changelog pages
npx @arabold/docs-mcp-server@latest scrape react https://react.dev/reference/react \
  --exclude-pattern "**/changelog*"

Output is a plain-text status line, e.g. Successfully scraped 42 pages. Progress updates appear on stderr during the run.

refresh

Re-scrape an existing library version, skipping unchanged pages via HTTP ETags.

npx @arabold/docs-mcp-server@latest refresh <library> [options]

Flag	Alias	Description
--version <ver>	-v	Version to refresh (omit for latest)
--embedding-model <model>		Embedding model configuration
--server-url <url>		Remote pipeline worker URL
--quiet		Suppress non-error diagnostics
--verbose		Enable debug logging

Example:

npx @arabold/docs-mcp-server@latest refresh react --version 19.0.0

The library and version must already be indexed. Use scrape for first-time indexing.

remove

Delete a library (or a specific version) from the index.

npx @arabold/docs-mcp-server@latest remove <library> [options]

Flag	Alias	Description
--version <ver>	-v	Specific version to remove (omit to remove latest)
--server-url <url>		Remote pipeline worker URL
--quiet		Suppress non-error diagnostics
--verbose		Enable debug logging

Example:

npx @arabold/docs-mcp-server@latest remove react --version 18.3.1

This is destructive and cannot be undone. Re-run scrape to re-index. Bulk deletes checkpoint the WAL without locking readers. Use compact when idle if the main store file is still large.

compact

Reclaim unused SQLite pages and truncate the WAL file so the store shrinks on disk. This takes an exclusive lock and may block searches until it finishes. VACUUM uses temporary files to keep memory usage low, so make sure there is enough free disk space for SQLite's temporary copy of the database.

npx @arabold/docs-mcp-server@latest compact [options]

Flag	Alias	Description
--force		Run VACUUM even when SQLite reports no free pages
--server-url <url>		Remote pipeline worker URL
--quiet		Suppress non-error diagnostics
--verbose		Enable debug logging

Example:

npx @arabold/docs-mcp-server@latest compact

Removing documentation does not shrink the main SQLite file. Bulk deletes only run a non-blocking WAL checkpoint. Run this command when the store is idle to VACUUM and reclaim free pages. Add --force only when you want to rewrite the store even if SQLite reports no free pages.

Output behaviour

All three commands write plain-text status messages to stdout and diagnostics to stderr. The global --output flag is accepted but has no effect because the output is plain text, not structured data.

In non-interactive sessions, diagnostics are suppressed by default. Use --verbose (or set LOG_LEVEL=INFO) to re-enable them. Use --quiet to suppress all non-error diagnostics regardless of session type.

Typical workflow

# 1. Index documentation for the first time
npx @arabold/docs-mcp-server@latest scrape react https://react.dev/reference/react --version 19.0.0

# 2. Later, refresh to pick up any changes
npx @arabold/docs-mcp-server@latest refresh react --version 19.0.0

# 3. Clean up old versions
npx @arabold/docs-mcp-server@latest remove react --version 18.3.1

# 4. Reclaim unused store space if needed
npx @arabold/docs-mcp-server@latest compact

Important notes

• Scraping can take time. Large documentation sites with hundreds of pages may run for several minutes. Use --max-pages and --max-depth to limit scope when you only need a subset.
• Local files must use the file:// URL scheme (e.g. file:///absolute/path/to/docs).
• scrape never overwrites by accident. Scraping a library version that already exists fails and leaves it untouched. Pass --replace to rebuild it from scratch, use refresh to update it in place, or pass --no-clean to add more documents to it.
• Library names keep their casing (React, Next.js Docs) and are matched case-insensitively, so react reaches React. No character in a name has a special meaning: use --version 18, not react@18.
• refresh only works on previously indexed content. It uses HTTP ETags to skip pages that have not changed, making it much faster than a full re-scrape.