Goodreads, from the command line

Goodreads has no public API any more, so getting structured data out of it usually means writing a scraper, guessing at selectors, and doing it again the next time a page changes. goodread does that part for you. It reads a public page, folds it into a record with real fields, and records where every number came from.

goodread book 2767052       # a printing, with its work, series and contributors
goodread work 2792775       # the thing itself, read through its best edition
goodread editions 2792775   # every printing, which is where the ISBNs are
goodread author 153394      # an author and the books their page lists

It talks to www.goodreads.com over plain HTTPS with no API key. The binary is pure Go with no runtime dependencies. Output is a table on a terminal and JSONL when piped, with json, csv, tsv, url, rdf, --fields and --template when you want something else.

What you can do with it

• Read a book, and know it is not a work. A work is what somebody wrote and a book is one printing of it. The Hunger Games is one work with 527 editions, each with its own ISBN, page count and rating population. book gives you a printing and work gives you the thing itself.
• Know where every number came from. Each record carries the surface it was read from and which rung of the extraction ladder answered for each field. goodread extraction prints the whole ladder.
• Build a graph, then query it offline. Everything read is folded into SQLite as nodes and edges over a closed set of twelve predicates. graph walks it, query runs SQL over it, find does full text search, and export writes JSONL or RDF.
• Crawl politely, and resume. crawl walks out from a set of seeds one connection at a time, and there is no flag to make it more. The frontier lives in the store, so an interrupted crawl continues where it stopped.
• Serve it to a model. goodread mcp is an MCP server over stdio with eleven read tools, and no search, shelf or reviews.
• Script it. Stable exit codes that say what went wrong rather than how the run ended, and an id command that classifies a URL without fetching.

What it will not read by default

goodread obeys robots.txt. Four surfaces are disallowed by Goodreads' rules and are not read unless you say so: the search page, shelves, reviews and /work/<id>. There is a --no-robots flag for a person who has decided it is their call, and it warns once, and the pace floor still applies.

The MCP server never reads a disallowed surface at all, and --no-robots has no effect on it. The flag's whole justification is a person deciding it is their call, and a model calling a tool is not that person deciding.

Independent and public-data only

goodread is an independent, open-source tool. It is not affiliated with, endorsed by, or sponsored by Goodreads or Amazon. It reads only public pages, at a polite default rate of one request a second at the very fastest and two seconds by default.

Where to go next

• New here? Start with the introduction for the mental model, then the quick start.
• Want to install it? See installation.
• Looking for a specific task? The guides cover books and works, the graph and the store, crawling, the MCP server, robots and output formats.
• Need every flag? The CLI reference is the full surface.