Skip to content

Privacy Policy

Last updated: 17 November 2026

AudioSilo Meta is a public, open source audiobook metadata database: this website, an API over the catalogue, and the whole catalogue published as downloadable SQLite releases. There are no user accounts and nothing to sign in to. We set no cookies and run no analytics, advertising, or tracking scripts of any kind.

Who we are

AudioSilo Meta ("the database") is published by KodeStar. Questions about this policy can be sent to kode@audiosilo.app.

The short version

  • No accounts, no sign-in, no cookies.
  • No analytics, advertising, or tracking scripts - we do not profile visitors.
  • No access logs: the web server keeps only an error log, used only to run the service and deal with abuse.
  • Your library export never leaves your device: /import parses it entirely in the browser.
  • Your watchlist stays in your browser - /watching sends nothing to the server. The optional AudioSilo Discord bot keeps only what you choose to send it.
  • The database holds bibliographic facts about published books and the people credited on them - public information about public creative work.

Server logs

The web server in front of this site keeps no access log of the requests it answers over HTTPS, which is every page and every API call: the requested URL, the time, your IP address and your browser's user agent are not written down. A plain http:// request, which is only redirected to HTTPS, can be.

It does keep an error log, which can record the IP address and the requested URL of a request that goes wrong, such as one that cannot reach the service while it is restarting. To apply a per-client rate limit, the service also holds a client's address in memory, not on disk, while it is making requests and normally for a few minutes after. All of this is used only to operate the service and to deal with abuse such as scraping floods or attacks. It is not used to profile anyone and it is not shared.

Search

Search queries go to this site's own API, on the same origin as the page you are reading. They are handled like any other request: no access log (see above), nothing more. There is no search history, because there is nobody to attach one to.

Language filter

The languages you choose under "Languages" in the header, and whether you have answered the one-line language suggestion, are stored only in your browser (in local storage on this device). To decide whether to offer that suggestion, the page reads the languages your browser already states; they are not sent anywhere. The languages you choose are sent only as the lang parameter on the search, latest additions and contribute-list requests they narrow.

Importing your library (/import)

The import page compares a library export against the database and shows you which of your books are catalogued and which are new. The export file is parsed entirely in your browser. The file itself never leaves your device, and nothing about your library is uploaded to us. Matching is done by asking the public API about identifiers and, for books it cannot match that way, author names - the same requests the search box makes.

Watching a series (/watching)

The series you watch and the marks you make on individual books - "I have this" and "not interested" - are stored only in your browser (in local storage on this device) and are never sent to the server - there is no account to attach them to. The page asks the public API only which books each series contains, the same request the series page makes, and clearing this site's data clears your watchlist, which is why the page offers a download of it.

To show a count of new releases next to "Watching" in the site header, this browser asks the public API for your own watched-series feed at most once an hour. That request carries only the series slugs - the same list a subscription URL carries - and its answer is kept in this browser until the next hourly check. If you are not watching anything, no such request is made.

A notification feed URL, and the calendar subscription URL that is the same feed in calendar form, carry the watched series slugs in the URL and are served statelessly without storing a subscription. Like any other request, a fetch of one is not written to an access log here. Whatever service you subscribe with - a feed reader, a calendar app, an automation service - sees that URL and the series it lists, and keeps whatever logs it keeps.

The one notification service we run ourselves is the bot on the AudioSilo Discord. If you send it your feed URL, it stores your Discord user ID, that URL, whether you have paused it, and the feed entries it has already sent you or is waiting to send, so it knows what is new. Send it stop to delete all of it; leaving the server removes it too. Your messages are handled by Discord under its own privacy policy.

The lookup assist (/add)

On the add-a-book page, searching for a title queries libex, a third-party Audible-metadata mirror, directly from your browser. That service therefore sees the text you searched for and your IP address, under its own terms. This happens on /add and nowhere else on the site.

Cover images

The database stores cover art as URLs, not as files - typically pointing at retailer content delivery networks. When a page shows covers, your browser fetches those images from the hosts that serve them, and those hosts see your IP address as they would for any image on the web.

Reports of unknown books

Where this optional feature is turned on, the service accepts reports of Audible ASINs the database does not hold yet. A report carries ASINs and nothing else we keep: the service stores each ASIN with a count and the first and last date it was reported, and stores no IP address with them. An address is held in memory for about a day after its last request to enforce a rate limit. These reports are never written to an access log when they arrive over HTTPS, which is how AudioSilo servers send them: the web server skips logging them on their own, whatever it does for the rest of the site. Its error log can still record the address of a request that goes wrong, such as a report that cannot reach the service while it is restarting.

Contributing

Contributions are made on GitHub, through issues and pull requests, under GitHub's own terms and privacy policy. Contributions are public and permanently recorded in the repository's history, and the data itself is published under open licences (CC0-1.0 for the factual core, CC BY-SA 4.0 for the community layer). A contributor's GitHub username is therefore part of the public record and cannot meaningfully be withdrawn later.

People in the database

The database contains bibliographic facts about published authors, narrators and other credited contributors: their names, the works they are credited on, and public identifiers. This is public information about public creative work - the same material a library catalogue or a bookshop listing carries - not private personal data. If you are credited in the database and something is wrong, or you have a concern about a record, please open an issue or email kode@audiosilo.app.

External links

Pages here link out to GitHub, Discord, audiosilo.app and other sites. Once you follow such a link you are on a service governed by its own privacy policy, not this one.

Changes to this policy

We may update this policy as the service evolves. Changes will be reflected by the "Last updated" date above and published at the same URL.

Contact

Questions about this privacy policy can be sent to kode@audiosilo.app.