TypeScript client for Kindle's unofficial API.
This package is rewrite and extension of https://github.com/Xetera/kindle-api.
It provides a simple TypeScript client for accessing Kindle's unofficial API.
Important
This library is not officially supported by Amazon / Kindle. Using this library might violate Kindle's Terms of Service. Use it at your own risk.
npm install kindle-api-kyThis library does depends on an external tls-client-api to proxy requests due to amazon's recent changes to their TLS fingerprinting. You'll need to run the server locally to be able to use this library. It's quite easy to set up and have one running in a few minutes and will save you tons of headache if you wanna do other kinds of scraping in the future.
This is the `config.dist.yml` file I'm using locally for testing:
The only thing I changed form the defaults is the api_auth_keys which you'll need to set as an environment variable TLS_SERVER_API_KEY.
env: dev
app_project: tls-client
app_family: tls-client
app_name: api
log:
handlers:
main:
formatter: console
level: info
timestamp_format: '15:04:05:000'
type: iowriter
writer: stdout
sentry:
dsn: ''
release: ''
tags:
project: tls-client
component: tls-client-api
api:
port: 8080
mode: release
health:
port: 8081
timeout:
read: 120s
write: 120s
idle: 120s
api_auth_keys: ['agentic-auth-key']
api_cors_allowed_origin_pattern: ''
api_cors_allowed_headers: ['X-API-KEY', 'X-API-VIEW', 'Content-Type']
api_cors_allowed_methods: ['POST', 'GET', 'PUT', 'DELETE']Amazon's login system is quite strict and the SMS 2FA makes automating logins difficult. Instead of trying to automate that with puppeteer and slow things down, we use 4 cookies that stay valid for an entire year.
TODO: One or more of these cookies is expiring every few minutes...
ubid-mainat-mainx-mainsession-id
You can grab these values directly by going on inspect element after loading read.amazon.com and copying the entire thing or just the select ones.
We also need a deviceToken for your kindle. You can grab this from the same network window as before on the getDeviceToken request that looks like:
Both of those identifiers should be the same.
import { KindleClient } from 'kindle-api-ky'
const kindle = new KindleClient({
cookies: 'ubid-main=xxx.xxxx ...',
deviceToken: '(your-device-token)'
tlsServerUrl: 'http://127.0.0.1:8080',
tlsServerApiKey: '(your-tls-server-api-key)'
})
// Initialize session and fetch initial list of books
await kindle.init()
console.log(kindle.books)
/*
[
{
"title": "Revelation Space (The Inhibitor Trilogy Book 1)",
"asin": "B0819W19WD",
"webReaderUrl": "https://read.amazon.com/?asin=B0819W19WD",
"productUrl": "https://m.media-amazon.com/images/I/41sMaof0iQL._SY400_.jpg",
"authors": [
"Alastair Reynolds"
],
"resourceType": "EBOOK",
"originType": "PURCHASE",
"mangaOrComicAsin": false
},
{
"title": "Dragon's Egg: A Novel (Del Rey Impact)",
"asin": "B004G8PJDA",
"webReaderUrl": "https://read.amazon.com/?asin=B004G8PJDA",
"productUrl": "https://m.media-amazon.com/images/I/51SEoLeSZuL._SY400_.jpg",
"authors": [
"Robert L. Forward"
],
"resourceType": "EBOOK",
"originType": "PURCHASE",
"mangaOrComicAsin": false
},
// ...
]
*/Instead of passing values directly to KindleClient, you can alternatively use env vars:
KINDLE_COOKIES='TODO'
KINDLE_DEVICE_TOKEN='TODO'
TLS_SERVER_URL='TODO'
TLS_SERVER_API_KEY='TODO'You can checkout this repo, pnpm install, set up .env, and run tsx bin/test.ts to test things locally.
const bookDetails = await kindle.getBookDetails(kindle.books[0]!.asin)
console.log(bookDetails)
/*
{
"title": "Revelation Space (The Inhibitor Trilogy Book 1)",
"asin": "B0819W19WD",
"authors": [
"Alastair Reynolds"
],
"bookType": "owned",
"formatVersion": "CR!WPPV87W8317H7FWJRF6JFMVE7SJY",
"mangaOrComicAsin": false,
"originType": "PURCHASE",
"productUrl": "https://m.media-amazon.com/images/I/41sMaof0iQL._SY400_.jpg",
"coverUrl": "https://m.media-amazon.com/images/I/41sMaof0iQL._SY400_.jpg",
"largeCoverUrl": "https://m.media-amazon.com/images/I/41sMaof0iQL.jpg",
"metadataUrl": "https://k4wyjmetadata.s3.amazonaws.com/books2/B0819W19WD/da38557c/CR%21WPPV87W8317H7FWJRF6JFMVE7SJY/book/YJmetadata.jsonp?X-Amz-Algorithm=AWS4-HMAC-SHA256&X-Amz-Date=20241004T063350Z&X-Amz-SignedHeaders=host&X-Amz-Expires=600&X-Amz-Credential=AKIAUHDZ6VO6DPMXT2XV%2F20241004%2Fus-east-1%2Fs3%2Faws4_request&X-Amz-Signature=e6cd53784e98d2ea5b8cea256c9403341be4ae103f31fabddd26cc173a0cd1f1",
"progress": {
"reportedOnDevice": "Travis's 2nd iPad",
"position": 163586,
"syncDate": "2024-10-04T06:28:56.000Z"
},
"webReaderUrl": "https://read.amazon.com/?asin=B0819W19WD",
"srl": 2393,
"percentageRead": 13.100000000000001,
"releaseDate": "21/04/2020",
"startPosition": 5,
"endPosition": 1250310,
"publisher": "Orbit"
}
*/Kindle uses heavy DRM and obfuscation for the actual book contents. We can, however get some rendering manifest data that is very useful. This method Returns a TAR file as a binary-encoded string. Unzipping the TAR file will result in about a dozen JSON files which specify different aspects of the book's pre-rendered content, layout, TOC, and metadata.
const manifestTar = await kindle.getBookContentManifest(kindle.books[0]!.asin)This library is not endorsed or supported by Amazon / Kindle. It is an unofficial library intended for educational purposes and personal use only. By using this library, you agree to not hold the author or contributors responsible for any consequences resulting from its usage.
MIT © Travis Fischer
This package is rewrite and extension of https://github.com/Xetera/kindle-api.
If you found this project interesting, consider following me on Twitter.