Skip to main content

MCP server for book metadata lookup — searches Google Books and Open Library with intelligent scoring

Project description

book-metadata-mcp

An MCP server that searches Google Books and Open Library to find book metadata, cover art, and publication info. Built for use with Claude Code, Claude Desktop, and any MCP-compatible client.

Features

  • Multi-source search — Queries Google Books and Open Library simultaneously, scores results by match quality
  • High-res covers — Google Books covers up to ~1280px wide (6 sizes), Open Library L-size covers
  • Smart scoring — Title/author matching with study guide detection (CliffsNotes, SparkNotes automatically deprioritized)
  • Accurate publication years — Cross-references both sources, picks the earliest credible year (avoids modern reprint dates)
  • Batch mode — Search up to 20 books at once
  • No API keys required — Works out of the box with free tiers

Tools

Tool Description
search_book Search by title, author, or ISBN. Returns scored results from both sources.
get_cover Find the best available cover image URL (prioritizes largest size).
get_metadata Get merged metadata: author, year, description, ISBN, subjects, page count.
download_cover Download cover image and save as JPEG. Requires Pillow.
bulk_search Search multiple books at once (up to 20 per batch).

Installation

Claude Code

claude mcp add book-metadata -- uvx book-metadata-mcp

Claude Desktop

Add to your claude_desktop_config.json:

{
  "mcpServers": {
    "book-metadata": {
      "command": "uvx",
      "args": ["book-metadata-mcp"]
    }
  }
}

From source

git clone https://github.com/colonylibrary/book-metadata-mcp.git
cd book-metadata-mcp
pip install -e ".[covers]"

Then register with Claude Code:

claude mcp add book-metadata -- book-metadata-mcp

Usage Examples

Once installed, the tools are available to your AI assistant:

"Search for The Great Gatsby by F. Scott Fitzgerald"

{
  "query": {"title": "The Great Gatsby", "author": "F. Scott Fitzgerald"},
  "result_count": 5,
  "results": [
    {
      "source": "open_library",
      "title": "The Great Gatsby",
      "authors": ["F. Scott Fitzgerald"],
      "published_date": "1920",
      "cover_url": "https://covers.openlibrary.org/b/id/10590366-L.jpg",
      "match_score": 95.0
    }
  ]
}

"Get the cover for Dune by Frank Herbert"

Returns URLs for the largest available cover from Google Books (up to extraLarge ~1280px) with fallback to Open Library.

"Look up metadata for these books: 1984, Brave New World, Fahrenheit 451"

Batch mode searches all three, returning title, author, year, ISBN, and cover URL for each.

Configuration

All configuration is optional via environment variables:

Variable Default Description
GOOGLE_BOOKS_API_KEY (none) Optional Google API key for higher quota (1,000/day without)
GOOGLE_DELAY 0.5 Seconds between Google API calls
OPENLIBRARY_DELAY 0.35 Seconds between Open Library API calls
GOOGLE_CB_THRESHOLD 3 Consecutive 429 failures before circuit breaker trips
GOOGLE_CB_COOLDOWN 60 Seconds to skip Google after circuit breaker trips
BOOK_MCP_USER_AGENT BookMetadataMCP/0.1.3 User-Agent for API requests

Bulk Usage / Rate Limiting

Google Books allows ~1,000 requests/day without an API key. For libraries larger than ~500 books, you'll hit rate limits. The server handles this automatically:

  1. Circuit breaker — After 3 consecutive 429 responses, Google is skipped for 60 seconds
  2. Graceful fallback — Open Library continues serving results during Google cooldown
  3. Auto-recovery — Google is retried after the cooldown period expires
  4. Zero errors — Rate limiting never causes failures, just temporary loss of Google-specific data

For the best bulk experience, set GOOGLE_BOOKS_API_KEY to a Google Cloud API key with Books API enabled (free tier: 1,000/day; paid: higher).

How Scoring Works

Each result is scored on a 0-100+ scale:

Factor Points Description
Title match 0-50 Exact match = 50, partial = proportional
Author match 0-30 Exact = 30, partial/substring = 20
Has cover 10 Cover image URL available
Has description 5 Book description available
Has ISBN 5 ISBN-10 or ISBN-13 available
Google source 2 Slight preference for cover quality
Study guide -60 CliffsNotes, SparkNotes, etc.
Inflated title -15 Result title much longer than query

Data Sources

Google Books API

  • Publisher-verified metadata
  • 6 cover image sizes (smallThumbnail through extraLarge)
  • Rich descriptions from publishers
  • Free: 1,000 requests/day without API key

Open Library API

  • Community-curated metadata
  • Original publication years (first_publish_year)
  • Extensive subject classification
  • Free: 3 requests/sec with User-Agent header

Stress Test Results

Tested against 1,920 audiobook titles (v0.1.3, fresh install, no API key):

Metric Result
Match rate 92.2% (1,752 / 1,901 processed)
Errors 0
High-confidence matches (>=70) 80% (1,523)
Has cover URL 85% (1,629)
Has ISBN 75% (1,444)
Has publication year 99.9% of matches
Speed 0.5-0.7 books/sec
Google 429 handling Circuit breaker tripped, graceful OL fallback

All 5 tools pass: search_book, get_cover, get_metadata, download_cover, bulk_search. Edge case tests: 26/26 passed (Unicode, short titles, empty inputs, special characters, ISBN lookup, bulk limits).

License

MIT

Project details


Download files

Download the file for your platform. If you're not sure which to choose, learn more about installing packages.

Source Distribution

book_metadata_mcp-0.1.3.tar.gz (12.4 kB view details)

Uploaded Source

Built Distribution

If you're not sure about the file name format, learn more about wheel file names.

book_metadata_mcp-0.1.3-py3-none-any.whl (12.9 kB view details)

Uploaded Python 3

File details

Details for the file book_metadata_mcp-0.1.3.tar.gz.

File metadata

  • Download URL: book_metadata_mcp-0.1.3.tar.gz
  • Upload date:
  • Size: 12.4 kB
  • Tags: Source
  • Uploaded using Trusted Publishing? No
  • Uploaded via: twine/6.2.0 CPython/3.12.7

File hashes

Hashes for book_metadata_mcp-0.1.3.tar.gz
Algorithm Hash digest
SHA256 abe0af9a91b3533d831d2df9b5f37702f2549fb813e61a9e42a6d38779d0d383
MD5 7e59c8a6d0b5d5fa8778f159384c1150
BLAKE2b-256 5f60e7852e408067ddb1d75f2ace4da75489c5ca6b7aec5d58a82db0cfe07873

See more details on using hashes here.

File details

Details for the file book_metadata_mcp-0.1.3-py3-none-any.whl.

File metadata

File hashes

Hashes for book_metadata_mcp-0.1.3-py3-none-any.whl
Algorithm Hash digest
SHA256 229f8fdc631febd2cdab022e4a123938a8f30ef97c8fe2bfa497ba53dc5188ca
MD5 a77342f277593f3480ce26bdde123d70
BLAKE2b-256 4ad8b937ef18758e75bf6b87569e0cfc02d1396d0013e38719efc20d42b42ef9

See more details on using hashes here.

Supported by

AWS Cloud computing and Security Sponsor Datadog Monitoring Depot Continuous Integration Fastly CDN Google Download Analytics Pingdom Monitoring Sentry Error logging StatusPage Status page