MIT Open Library MCP. Find titles, editions, and bibliographies from 20M+ records.

You're connecting your AI client directly to the Open Library catalog—that’s over 20 million books and all their associated metadata. This isn't a simple search box; it's a deep dive engine for researchers. You gotta use this server when you need academic depth, not just what Google Books throws at ya.

We built out tools that let your agent pull structured data on authors, titles, and every damn edition they ever printed.

When you start up a search, you can hit the catalog with specific identifiers. Need to find a book by its ISBN-10 or ISBN-13? Use search_by_isbn and you instantly get core details like the title, publication date, page count, and cover image. If you know the exact name of the work, run search_by_title.

You can also narrow your search dramatically using specific criteria. Want to filter results only for books published by 'MIT Press' or 'O’Reilly Media'? Use search_by_publisher or search_by_subject, plugging in academic codes like 'computer_science' or 'quantum_physics'. Need something in Chinese? You use search_by_language with the ISO code. If you want to see what people are talking about right now, check out search_trending_subjects.

And if your research is focused on finding free material, running search_full_text makes sure you only get books confirmed to have readable full text online.

If you're starting from an author's name—say, a common name like 'Isaac Newton'—you use search_by_author to look up their bibliographies. You can also search for authors using their precise Open Library key with get_author, which pulls the full profile and bio. Once you have that master list of works, get_author_works returns all titles associated with them, along with covers and primary subjects.

But here's where it gets complex—and useful. If an author has written a massive body of work, you can track down every single version of it. You start by using get_work, which pulls general details, including subject matter and historical context for the original title. From that master record, you use get_work_editions to list every known variation—every translation, format change, or re-release.

If you need the super specific details on one of those versions—like who published it, what its exact page count is, and which ISBNs it used—you call get_edition using the edition key.

For general searches that require a starting point, you can also run search_by_subject, or if you know the author's name but not their specific works, you use search_by_isbn. For finding recent material, just fire up search_recent to see what’s been added or updated in the catalog. The structure is designed so your agent doesn't hit a dead end; it links identifiers right through the entire lifecycle of a book.

Your AI client sends the key parameters—an ISBN, an author name, a subject code—and the MCP Server executes the call. It pulls back structured JSON containing title details, related editions, subjects, and all the necessary metadata for your agent to process without having to stitch together disparate data points from multiple sources.

You're getting organized academic power here. You don't just find a book; you trace its history, see every format it came in, and track down the specific version published by a certain house.