# LibraryThing MCP > LibraryThing MCP connects your AI agent to a massive bibliographic database. Pull exact book metadata, map all ISBNs across physical and digital editions, and check how thoroughly a title is cataloged. You get raw publication data, community ratings, and coverage scores without writing custom scrapers. ## Overview - **Category:** knowledge-management - **Price:** Free - **Tags:** books, isbn, bibliography, library, reading ## Description Look, if you are building an app that relies on book data, you know the pain. You either pay for a massive commercial API or spend weeks writing fragile web scrapers that break when a site changes its layout. This MCP gives you a direct line to the LibraryThing database. You just ask your agent to find a book, and it pulls the exact metadata, community ratings, and publication history. Because Vinkius hosts this connection, you skip the boilerplate of setting up HTTP clients and parsing messy HTML. You just plug it into your preferred client and start querying. It handles the messy part of bibliographic data. Books have dozens of editions. A single title might have a hardcover, a paperback, an audio version, and five international releases. This tool maps all those variations back to a single canonical work. You get a clear picture of what actually exists in the real world, plus a coverage score that tells you how well the community has cataloged it. It is basically a shortcut to clean, normalized book data. ## Tools ### get_book_coverage Returns a zero-to-one score showing how completely the community has cataloged a specific book. ### get_work Pulls detailed metadata like author, review counts, and member totals for a specific book record. ### thing_isbn Finds every ISBN for all physical and digital editions of a single title. ### what_work Translates a basic book query into the internal ID required for deeper lookups. ## Prompt Examples **Prompt:** ``` Look up the book with ISBN 978-0-13-468599-1. ``` **Response:** ``` Found it! 'The Pragmatic Programmer' by David Thomas and Andrew Hunt. Publisher: Addison-Wesley (2019, 2nd edition). Community rating: 4.2/5 with 1,847 members. Tagged: programming, software-engineering, best-practices. ``` ## Capabilities ### Map book editions Group every physical and digital format of a single title under one canonical record. ### Fetch publication details Pull exact author, publisher, and release date information for any known book. ### Calculate catalog coverage Get a zero-to-one score showing how completely the community has documented a specific title. ### Extract community metrics Grab member counts, review totals, and user-generated tags for any work. ### Resolve work identifiers Translate a messy book query into a stable internal ID for downstream lookups. ## Use Cases ### Fixing duplicate book records A user submits an ISBN for a French paperback, but your system already has the English hardcover. You use thing_isbn to find all variations, then use what_work to prove they share the same canonical ID and merge the records. ### Filtering out bad data for a recommendation engine You need high-quality metadata for a new feature. You run a batch of titles through get_book_coverage and drop anything with a score below 0.8, ensuring your model only trains on well-documented books. ### Building a reading list app A user asks for details on a specific sci-fi novel. Your agent uses what_work to get the ID, then calls get_work to pull the author, publication year, and community tags to display on the user's dashboard. ### Auditing a digital library catalog A librarian notices missing metadata for a collection of indie books. They use get_book_coverage to identify which titles lack community documentation and flag them for manual review. ## Benefits - Stop guessing which edition a user meant. The thing_isbn tool maps every paperback, hardcover, and audio format back to a single canonical record so your database stays clean. - Know exactly how well a book is documented before you rely on its data. The get_book_coverage tool returns a clear score so you can filter out poorly cataloged titles. - Skip the messy ID translation step. You just ask for a book, and the what_work tool instantly resolves it to the exact internal ID needed for deeper queries. - Get real community sentiment without scraping review sites. The get_work tool pulls member counts, review totals, and user tags directly from the LibraryThing database. - Eliminate the boilerplate of building HTTP clients. You connect this MCP once, and your agent handles all the API calls and JSON parsing automatically. ## How It Works The bottom line is you get instant access to a massive, community-maintained book database without writing a single line of parser code. 1. Subscribe to the MCP and generate a LibraryThing Developer Key from their services page. 2. Connect your AI client using the provided configuration and paste in your API key. 3. Ask your agent to look up a book by ISBN or title to start pulling bibliographic data. ## Frequently Asked Questions **How do I use the LibraryThing MCP get_work tool?** You first need the internal work ID. Use the what_work tool to translate your book query into an ID, then pass that ID to get_work to pull the full metadata. **Does the LibraryThing MCP thing_isbn tool require an API key?** Yes, you need to register for a LibraryThing Developer Key to use the MCP. Once you add that key to your client configuration, tools like thing_isbn will work without extra setup. **What does the LibraryThing MCP get_book_coverage score mean?** It is a zero-to-one rating showing how completely the community has cataloged a book. A score near one means the title has extensive metadata, while a low score means it is barely documented. **Can the LibraryThing MCP what_work tool find audiobooks?** Yes. It resolves the canonical work ID for a title regardless of the format. Once you have that ID, you can use thing_isbn to find the specific ISBNs for the audio, paperback, and hardcover editions. **How many tools are in the LibraryThing MCP?** There are four tools. You get what_work to find IDs, get_work for detailed metadata, thing_isbn to map editions, and get_book_coverage to check data quality. **How does the LibraryThing MCP thing_isbn tool handle international editions?** It returns all ISBNs for a specific title across different regions. This includes paperback, hardcover, and localized print runs so your agent can map global inventory. **What happens if the LibraryThing MCP what_work tool cannot find a matching book?** It returns a null identifier. Your AI client should handle this by prompting you for an ISBN or alternate title spelling before making another request. **Can I use the LibraryThing MCP get_work tool without a work ID?** No. You must pass a valid work ID to get detailed metadata. Run the what_work tool first to resolve a title or ISBN into the required numeric identifier. **What information can I get from an ISBN lookup?** An ISBN lookup returns the book title, author(s), publisher, publication date, page count, language, community ratings, and tags — all from the LibraryThing database.