📖 Open Library Scraper: Books, Ratings & ISBN Data

Open Library Scraper is an Apify actor that exports book, edition, work, author and subject data from openlibrary.org. Search by title, author, subject, publisher, language, ebook access or year and get ratings, want-to-read counts and edition counts per book. Look up ISBNs, work IDs, authors and subjects. Up to 91 fields, no login, no API key. Export CSV, Excel, JSON or XML. Books cost $13 per 1,000 rows on the base tier, with no start fee. Last updated 2026-10-11.

Open Library Scraper turns the Open Library catalog into clean datasets. A live check on 2026-10-03 found 48,254 results for the query "dune" and 1,407 works for the author Frank Herbert. Dune alone shows 446 ratings, 3,418 want-to-read and 155 editions.

📚 What does Open Library Scraper do?

• Search books by keyword with optional author, subject, publisher, language (3-letter code or English name, comma = OR), ebook access, minimum ratings count and first-publication year range. A query needs at least 3 characters, or browse with filters alone.
• Look up ISBNs (one row per edition) and work IDs, following merged works.
• Author mode (names, IDs or URLs) and subject mode (names or slugs such as place:london).
• Skip duplicates and never charge for error, not-found or empty rows.

🗂️ What data can you extract from Open Library?

The dataset has 91 fields and four views (Books, Authors, Subjects, Editions). Each row has a recordType: book, edition, work, author or subject.

Record type	Main fields
Book (search)	title, authors, firstPublishYear, ratingsAverage, ratingsCount, wantToRead, currentlyReading, alreadyRead, editionCount, languages, ebookAccess, subjects, publishers, isbn, coverUrl, url
Edition (ISBN)	isbn10, isbn13, title, publishers, publishDate, pages, physicalFormat, workId
Work (work ID)	title, description, subjects, ratingsAverage, ratingsBreakdown (1 to 5 stars), reading counts
Author	authorId, name, birthDate, workCount, topWork, topSubjects, plus bio, links, remoteIds (Wikidata, VIAF, ISNI) with details
Subject	subjectKey, name, workCount, ebookCount, relatedSubjects, topAuthors, topPublishers

New columns are left out of a row when empty. Every row follows one fixed order of 91 columns and has scrapedAt.

🧾 Sample output

A real book row from a cloud run (author filter Frank Herbert; newer columns not shown):

{
  "recordType": "book",
  "workId": "OL893414W",
  "title": "Dune",
  "authors": ["Frank Herbert"],
  "firstPublishYear": 1965,
  "ratingsAverage": 4.3,
  "ratingsCount": 446,
  "wantToRead": 3418,
  "alreadyRead": 727,
  "editionCount": 155,
  "subjects": ["Dune (Imaginary place)", "Fiction", "Science fiction"],
  "isbn": "9780441013593",
  "coverUrl": "https://covers.openlibrary.org/b/id/11481354-M.jpg",
  "url": "https://openlibrary.org/works/OL893414W",
  "scrapedAt": "2026-10-04T04:18:23.444Z"
}

💰 How much does it cost to scrape Open Library?

Pay per event, charged only for rows saved to your dataset. Prices are the base (free plan) tier; paid plans get lower rates (read 2026-10-10).

Event	What you get	Price (base tier)
book-record	One searched book with ratings and reading stats	$0.013 ($13 per 1,000); $0.0125 Bronze, $0.012 Silver, $0.01 Gold and above
edition-record, work-record, author-record, subject-record	One ISBN edition, work, author or subject	$0.01 on every tier

Worked examples at the base tier: 10 searched books cost $0.13, 100 cost $1.30 and 1,000 cost $13.00. Looking up 1,000 ISBN editions or work details costs $10.00. 100 author rows cost 100 x $0.01 = $1.00, and 100 subject rows cost $1.00.

There is no start fee. You are not charged for error rows, "ISBN not found" rows, duplicates, empty searches or failed runs. A maximum cost per run is respected.

🚀 How to scrape Open Library (3 steps)

1. Click Try for free.
2. Enter a query or filters, or paste ISBNs, work IDs, authors or subjects, and set Max Items.
3. Click Start and download the dataset.

⚙️ Input

Field	Type	Description
searchQuery	string	Title or keywords, at least 3 characters. Optional
author / subject / publisher / language	string	Filters on the book search; language is a code or English name, comma = OR (eng,French)
ebooksOnly / ebookAccess	boolean / enum	Full-text ebooks only; or any, public, borrowable, printdisabled, no_ebook
yearFrom / yearTo, minRatingsCount	integer	Year range, minimum ratings
sortBy	enum	relevance, rating, new, old, editions, want_to_read, trending
includeWorkDetails, coverSize	boolean, enum	Work details on book rows (off by default); cover size S, M or L
isbns, workIds	string list	ISBN-10/13; work IDs such as OL893414W
authorSearch, authorIds, authorSortBy, includeAuthorDetails	string, list, enum, boolean	Search authors by name or give IDs (OL79034A) or URLs; optional biography, links, Wikidata/VIAF/ISNI, photo
subjectNames	string list	Names, slugs (place:london) or subject URLs
maxItems	integer	Row limit, default 10. Free users are capped at 10

At least one query, filter, ISBN, work ID, author input or subject is required. A search returns up to 100,000 rows per run; all modes share maxItems. An empty or invalid input ends as SUCCEEDED with one free error row.

{
  "searchQuery": "dune",
  "language": "eng",
  "ebooksOnly": true,
  "maxItems": 5
}

📦 Output

Results go to the default dataset with four views: Overview (books and works), Authors, Subjects and Editions. A search that matches nothing succeeds with "No results found" and charges nothing; if Open Library errors and zero rows are delivered, the run fails.

⚖️ Open Library Scraper vs alternatives

Store pages read on 2026-10-10. "Not stated" means the listing does not show it.

	This actor	parseforge/open-library-scraper	eins332570/book-scraper
Price per row	$0.013 Free, $0.01 Gold, no start fee	$0.012 plus $0.16 per run start	$0.005
Language and ebook filters	Yes, codes or names; access type	Ebooks toggle; language not stated	Language code, search only
Authors, subjects, ISBN, work IDs	Yes, all	Authors and subjects by search type	ISBN and work IDs
Reading lists, search inside	No	Yes, search types	Not stated in their listing
Fields	91	Not stated in their listing	19

parseforge is cheaper for small runs (10 rows: $0.13 here against $0.28 there) and about even at 100 rows ($1.30 against $1.36). At 1,000 rows on the base tier we are about 7% dearer ($13.00 against $12.16). logiover/openlibrary-scraper ($0.003 per row) and bovi/openlibrary-scraper ($0.001) are cheaper per row and return fewer fields; neither lists a language or ebook filter. Choose a cheaper actor for bare search rows, this one for reading counts, filters, ISBN editions and author or subject records.

💡 Use cases

• Listing enrichment and apps: add pages, publisher, cover and ratings to every ISBN.
• Ebook catalogues: list borrowable ebooks in one language.

🔌 Run via API, schedule and integrations

curl -X POST "https://api.apify.com/v2/acts/RecordsData~open-library-books-scraper/runs?token=YOUR_TOKEN" \
  -H "Content-Type: application/json" \
  -d '{"searchQuery":"dune","language":"eng","ebooksOnly":true,"maxItems":10}'

Use the API, schedule runs, and send the dataset to Zapier, Make, n8n or Google Sheets.

🤖 Open Library Scraper MCP server for AI agents

AI agents can call Open Library Scraper through the Apify MCP server at https://mcp.apify.com?tools=RecordsData/open-library-books-scraper. The URL loads only this Actor as a tool (Claude Desktop, Cursor, VS Code):

{
  "mcpServers": {
    "apify": {
      "url": "https://mcp.apify.com?tools=RecordsData/open-library-books-scraper"
    }
  }
}

Example agent prompt:

Use the Open Library Scraper to list 10 English books with a public ebook about the Roman Empire, sorted by want-to-read, then give me each title with its ratingsAverage, wantToRead and editionCount.

Tool description: choose Open Library Scraper when an agent needs book metadata from Open Library: ratings and reader-demand counts for searched books, language and ebook-access filtering, publisher and page data for an ISBN, the description and rating breakdown for a work ID, or an author or subject record.

⚖️ Is it legal to scrape Open Library?

The actor reads Open Library's public API, which serves open catalog data, with an identified user agent and polite delays. It collects book metadata only, not personal data. Follow Open Library's terms and cover image licenses.

❓ Frequently asked questions

What is the difference between a work and an edition in Open Library Scraper?

A work is the book as a title, such as Dune, and an edition is one of its printings, such as the 155 editions of Dune. The scraper returns works for keyword, author or subject searches, and edition rows for ISBN lookups. Work IDs return full work details.

Where do the ratings in Open Library Scraper come from?

They come from Open Library's own reader ratings on openlibrary.org. The scraper returns the average rating and the number of ratings for each book, plus want-to-read, currently-reading and already-read counts, so you can rank titles by reader demand. Work lookups also return the 1 to 5 star breakdown.

Does Open Library Scraper need an API key or login?

No. It reads the public Open Library API, which serves open catalog data, so there is no login, no API key and no cookie to set up. You only need an Apify account to run the actor, and the free plan works for small tests.

Why did Open Library Scraper return 0 results?

The search matched nothing on Open Library, for example with a very narrow combination of filters. Try fewer filters. The run ends as succeeded with the message "No results found", and nothing is charged for the empty search, so you can test narrow filters freely.

What happens if I send an empty or invalid input to Open Library Scraper?

Wrong enum values or types are rejected with a 400 error before any run starts. Other mistakes, such as a query under 3 characters, an unknown language, yearFrom later than yearTo or no input, end the run as SUCCEEDED with one free error row and no charge.

Why is there a row with an error field in Open Library Scraper output?

An ISBN, work ID, author or subject was not found on Open Library, or Open Library returned an error for that lookup. The scraper keeps such a row so you can see which input failed, and never charges for it. A long list of wrong ids stops after max(25, min(maxItems, 1000)) failures.

How many rows can I get from Open Library Scraper?

Free users get 10 rows per run. Paid users can request up to 1,000,000 rows through maxItems, and a single book search returns up to 100,000 rows. Paid users can raise maxItems up to 1,000,000. Split larger harvests into several searches by author, subject, language or year.

Can I get cover images from Open Library Scraper?

Yes. Every row for which Open Library has a cover includes a direct covers.openlibrary.org image URL in coverUrl, next to the canonical link in url. Pick the size with coverSize. Only links are returned, so check cover licenses before republishing the images.

Can I limit my spend with Open Library Scraper?

Yes. Set Max Items to cap the number of rows, or set the maximum cost per run in Apify. The scraper stops cleanly when the limit is reached and you pay only for rows saved to the dataset. A book row costs $0.013 and other rows $0.01 at the base tier.

How do I filter Open Library books by language?

Set language to a 3-letter Open Library code or an English name, for example eng or French. Separate several with commas to match any of them. The filter keeps works with an edition in that language. An unknown language ends with a free error row.

How do I get only ebooks with Open Library Scraper?

Turn on ebooksOnly to keep books that have a full-text ebook on Open Library or the Internet Archive. For finer control set ebookAccess to public, borrowable, printdisabled or no_ebook. Rows carry ebookAccess, hasFulltext and ebookCount, so you can also check access in the dataset.

Can Open Library Scraper return author data?

Yes. Use authorSearch with a name, or authorIds with Open Library author IDs or URLs. Author rows have birth and death dates, work count, top work, top subjects, alternate names and reader counts. includeAuthorDetails adds biography, links, Wikidata, VIAF and ISNI ids and a photo.

Can Open Library Scraper return subject data?

Yes. Give subjectNames as names or slugs such as science fiction, place:london or time:20th_century, or as a subject URL. Each subject gives one row with work count, ebook count, related subjects, places, people, times, top authors and top publishers; books under a subject come from the subject filter.

Can I browse Open Library without a search query?

Yes. A query is optional. Filters alone, such as language, publisher, ebooksOnly, ebookAccess, a year range or minRatingsCount, are a valid run. If you do give a query, it needs at least 3 characters; a shorter one ends the run with a free error row and no charge.

Can I get 100,000 rows from one Open Library search?

Yes, if your plan allows it. One book search returns up to 100,000 rows per run, and maxItems sets the limit (free users get 10). Large runs take longer because Open Library answers deep result pages slowly, so set Max Items or a maximum cost per run to keep them bounded.

What does includeWorkDetails do?

It adds description, links and firstPublishDate to each book row. That needs one extra Open Library request per row, so runs are slower, and it is off by default. There is no extra charge, and you only need it on book search rows.

Why do very large subjects like history return only a work count?

Open Library sometimes answers a very large subject, such as history or poetry, with a bare record after about 11 seconds. The scraper retries once, then counts the works through the search index and returns the work count with the other subject columns left out.

Did the column order change in Open Library Scraper?

Yes. Every row now follows one fixed order of 91 columns, so columns that existed before moved position in CSV and Excel. Names and meanings did not change. If a script reads columns by position, check it once or switch to column names.

Can Open Library Scraper return reading lists or search inside books?

No. Open Library reading lists are user-generated, and the search-inside endpoint was failing with HTTP 500 when this version was built, so neither is part of this actor, and no data for them is returned. Books, editions, works, authors and subjects are covered.

🔗 Want more data? Other RecordsData scrapers

Wikipedia Articles Scraper, Hacker News Search Scraper and PubMed Articles Scraper.

🛟 Support

Issues tab or email contact.punkrecordsdata@gmail.com.

Last updated: 2026-10-11