Goodreads, the Amazon-owned book community with more than 150 million members reported in 2023, serves its public book, author and list pages to almost anyone. On 2 October 2026 we sent nineteen requests from six countries through residential exits: every plain HTTP fetch loaded on the first try, with the full text and ratings in the HTML. But the public API has taken no new keys since December 2020, and the Terms of Use exclude robots and data mining. So a proxy does not unlock anything here; it is for seeing a page from another country, not for collecting the catalog.
What people search for
Google has no autocomplete for "goodreads proxies". On the US results page the word is read literally: the top results are Goodreads pages for a young-adult novel called Proxy and two essay collections called Proxies, plus a 2021 blog post about using proxy data to spot fake reviews and a dev.to scraping tutorial. The related searches are "Goodreads proxies free", "Goodreads proxies list" and "Alex london goodreads".
The real intent shows up in the neighbouring queries. "goodreads scraper" completes to github, api, review scraper, web scraper, data scraper and python. "goodreads api" completes to key, access, alternative, documentation, free, reddit and "2025". In other words, people want book data, they found out the API is gone, and they are wondering whether proxies are the workaround. The answer depends on three things this article measures: what Goodreads actually serves, what it costs in traffic, and what its rules say.
The Goodreads API: retired, and verifiable
Goodreads opened a public developer API in October 2010. On 8 December 2020 it put a banner on its API pages saying it was "no longer issuing new developer keys" and planned to retire the tools; existing keys were deactivated after 30 days of inactivity, according to developers in the official Goodreads Developers group, and the Goodreads Help Center has an article titled "Why did my API key stop working?" that repeats the date.
We checked what is left. On 2 October 2026 the old /api address redirected to the Goodreads homepage, 55,796 bytes, carrying the message "You are not authorized to perform this action." There is no documentation page any more, robots.txt disallows /api to every crawler, and no official replacement exists. The site does still answer its own search box with a small JSON response (5,330 bytes for five results in our test), but that endpoint is undocumented, belongs to the website, and comes with no licence to use it; the same terms apply to it as to every page.
If what you need is book metadata (titles, authors, ISBNs, editions, subjects), an open catalog is the cleaner route. Our ../../collectors/book-data-api/ collector reads Open Library, the Internet Archive's community catalog, which is built to be reused.
What Goodreads serves to a proxied client
Goodreads sits behind Amazon CloudFront. We picked public pages of famous books and authors: the Harry Potter and the Sorcerer's Stone book page, J.K. Rowling's author page and the Best Books Ever list. No logins, no user profiles, no review pages.
| Request (2 October 2026) | Exit | Result | Bytes | Time |
|---|---|---|---|---|
| Book page, plain HTTP | US | 200, 1st attempt, 18,071 words | 733,870 | 5.2 s |
| Same book page, plain HTTP | UK | 200, 1st attempt | 732,483 | 4.3 s |
| Same book page, plain HTTP | Japan | 200, 1st attempt | 732,400 | 7.7 s |
| Same book page, plain HTTP | Brazil | 200, 1st attempt | 732,984 | 7.1 s |
| Same book page, rendered | US | 200, 18,076 words | 2,101,565 | 31.2 s |
| Book page, no-JS vs rendered audit | US, Germany | 200 both ways, same title, description, h1 and canonical | n/a | 17.8 to 18.3 s |
| Author page, plain HTTP | US | 200, 1st attempt, 4,548 words | 181,699 | 2.7 s |
| Same author page, plain HTTP | India | 200, 1st attempt | 182,154 | 9.4 s |
| Same author page, rendered | US | 200, 4,553 words | 1,301,923 | 54.5 s |
| Author page, no-JS vs rendered audit | UK | 200 both ways, 2,132 words both | n/a | 20.5 s |
| Best Books Ever list, plain HTTP | US | 200, 1st attempt | 786,737 | 10.0 s |
| Same list, rendered | Germany | Timed out | n/a | n/a |
| robots.txt | US | 200, 1st attempt | 2,647 | 1.3 s |
| Terms of Use | US | 200, 1st attempt | 74,157 | 2.4 s |
| Old /api page | US | Redirect to homepage, "not authorized" | 55,796 | 14.5 s |
Three things stand out. First, nothing was challenged: twelve plain HTTP fetches from six countries, twelve 200s on the first attempt, no interstitial and no captcha page. Second, the browser adds nothing to the content. The rendered book page had 18,076 words against 18,071 over plain HTTP, and the audits found the same title, description, heading and canonical with and without JavaScript. Third, the browser adds a lot of traffic: 2.9 times the bytes on the book page and 7.2 times on the author page, and the slowest render took 54.5 seconds.
The book page is also well structured. Its HTML carries a schema.org Book block with the ISBN, page count, awards and an aggregate rating (4.47 from 11,848,368 ratings and 207,473 reviews on the day). The canonical URL of this American edition points to the British Philosopher's Stone edition, so edition IDs and canonical IDs are not the same thing on Goodreads. The author page has no structured data block. You can run the same kind of check on any site with our ../../tools/waf-detector/.
What Goodreads' rules say
This is the part that decides the answer. Section 1 of the Goodreads Terms of Use grants a limited licence for "personal and non-commercial use" and then lists what that licence does not include. Among the exclusions:
- "any collection and use of any book listings, descriptions, reviews or other material included in the Service";
- "any use of data mining, robots, or similar data gathering and extraction tools";
- any resale, commercial use or derivative use of the service or its contents, and any copying for a commercial purpose without express written consent.
The robots.txt is narrower than the terms but points the same way. It blocks OpenAI's GPTBot and Common Crawl's CCBot from the whole site, asks Bingbot for a five-second crawl delay, and for all other user agents disallows /api, /search, /book/reviews/, /review/show, /review/list, /work (except edition and quote pages), shelves and the RSS feeds. Book, author and list pages are not disallowed, which is why search engines index them. You can paste the file and a path into our ../../tools/robots-txt-tester/ to see the verdict for your own crawler.
Put together: robots.txt does not forbid a crawler from reading a book page, but the terms do not license collecting book listings, descriptions or reviews, or using robots at all, and their terms forbid commercial reuse without written consent. A proxy does not change any of that. If you need Goodreads data for a product or research at scale, the route is a written agreement with Goodreads, or an open catalog for metadata.
What a proxy is actually useful for on Goodreads
There are legitimate jobs, and most of them are small and done by people in a browser:
| Job | Proxy type | Notes |
|---|---|---|
| An author or publisher checking how their own book page looks to readers in another country | Residential, country-targeted, in a normal browser | Every country we tried got the same page; worth confirming for your own title |
| Checking that buy links, giveaways and ads on your book page show correctly by market | Residential, country-targeted | Ad and retailer links are the parts most likely to differ |
| Running a publisher or author account from a stable office address | Static ISP, one account per address | Consistency, not scale |
| Watching the terms or robots.txt for changes | Residential, plain HTTP | Small pages: 74,157 and 2,647 bytes |
| Collecting books, ratings or reviews at scale | None | Not licensed by the terms without written consent; use an open catalog for metadata |
We used ../../residential-proxies/ for every request above. We did not test mobile or datacenter exits on Goodreads in this run, so we make no claim about them. For a single publisher account, ../../isp-proxies/ gives you the same address every session.
What changes by country
Almost nothing. Goodreads is an English-language site, and the book page came back with the same title, description and canonical from the US, the UK, Germany, Japan and Brazil. The byte counts differed by at most 1,470 bytes (732,400 to 733,870), which is per-request markup, not different content. The author page from India was 455 bytes heavier than from the US. What did change is speed: plain HTTP took 4.3 seconds from the UK and 7.7 from Japan for the same book page, and 2.7 seconds from the US against 9.4 from India for the author page. That is distance to the nearest CloudFront edge, not a different site.
So for a geo-check, Goodreads itself is not the variable. The things that can vary by market are the retailer buttons and ads around the page, which is what an author or publisher would actually be checking.
What the traffic costs
Page weight is the cost driver on residential traffic, so it is worth knowing before you plan anything. At the Residential Basic rate of $0.80 per GB, with the bytes we measured:
| Page | Bytes per load | 1,000 loads | Cost at $0.80/GB |
|---|---|---|---|
| Book page, plain HTTP | 733,870 | 0.734 GB | $0.59 |
| Book page, rendered | 2,101,565 | 2.10 GB | $1.68 |
| Author page, plain HTTP | 181,699 | 0.182 GB | $0.15 |
| Author page, rendered | 1,301,923 | 1.30 GB | $1.04 |
| List page, plain HTTP | 786,737 | 0.787 GB | $0.63 |
The worked example: 1,000 book page loads × 733,870 bytes = 733,870,000 bytes, or 0.734 GB; 0.734 × $0.80 = $0.59. Rendering the same thousand pages moves 2.10 GB and costs $1.68, for exactly the same text. On Goodreads a browser is pure overhead. For the realistic jobs above (one author checking a page by hand from three countries) the traffic is a few megabytes and the cost rounds to zero.
If you are looking at other platforms where public collection is allowed, our ../../web-scraping-api/ returns structured results and bills only successful requests. For the bigger book retailer behind Goodreads, see our post on ../amazon-proxies/, and for community discussion data, ../reddit-proxies/.