LibriVox
librivox.org · Debrid · 1 record · JSON
LibriVox Debrid
★ qualifiesanonymousno cardDper-ip-shaping
| Free tier | Free public-domain audiobooks, read by volunteers, with a JSON/XML catalogue API. The API caps a single response at 500 records ('Can no longer be set above 500', added 2026-09-26) and pages with offset. unit unlimited; metered on none; The catalogue size is not published; the highest project id returned by sort_order=desc was 23017, which bounds it near 23,000 projects. |
|---|---|
| Price | USD 0 |
| Per-object cap | not applicable: one MP3 per section, one ZIP per book. The ZIP is served by archive.org, so the object size is the Internet Archive item's size. |
| Retention | permanent: public-domain recordings 'donated to the public domain', with the audio stored as Internet Archive items |
| Egress | Metadata is free and unmetered from the API. The audio is not served by librivox.org: each record's url_zip_file points at archive.org/compress/<item>/..., so audio bytes are Internet Archive egress and inherit its per-IP shaping. free_allowance: no byte allowance published |
| Identity | anonymous No key and no account: the API answered an anonymous GET with 200 and 3,084 bytes for limit=2. |
| Limits | |
| Binding | per-ip-shaping The API is metadata and would take minutes to walk at 500 records a request; the bytes come from archive.org, so the Internet Archive's per-IP ceiling is what decides a TiB-scale pull. |
| Interface | rest, cli GET https://librivox.org/api/feed/audiobooks with id, since, author, title, genre, fields, extended, coverart, limit, offset, format, sort_order. Parameters may be path segments instead of query strings. |
| Tooling | curl/python-requests; any HTTP client. No SDK and no rclone backend is needed because the metadata is a plain JSON API and the audio is a plain archive.org URL. |
| Composes with | Harvest the catalogue through the API (a few hundred requests), then pull the MP3s from archive.org - the same egress path and the same shaping as any other Internet Archive item. |
| Reliability | The API changed in 2026-09-26 to cap limit at 500; older clients that asked for more are silently clamped rather than rejected. |
| Notes | The interesting part is where the bytes are not: librivox.org serves a small JSON catalogue and archive.org serves the audio. A plan that expects a fast audiobook mirror from this API alone will move metadata and no media. |
| Evidence rung | documented The API contract is the vendor's own page and was exercised here for two records; the audio path is read from the record's url_zip_file field, not downloaded. Downloading one ZIP would raise it. |
| Checked | 2026-10-24 (live); stale on 2026-11-23 |
| Parsed numbers |
|
Evidence (4)
| date | source | what it proved |
|---|---|---|
| 2026-10-23 | librivox.org/api/info | 'limit - ... Defaults to 50. Can no longer be set above 500', added 2026-09-26, and the 'add a delay between your requests' guidance |
| 2026-10-23 | librivox.org/pages/about-librivox/ | 'Librivox donates its recordings to the public domain' and the non-commercial, ad-free terms |
| 2026-10-23 | probe:https://librivox.org/api/feed/audiobooks?limit=2&format=json&sort_order=desc | 200, 3,084 bytes of JSON; highest project id 23017, which bounds the catalogue near 23,000 projects |
| 2026-10-23 | probe:https://librivox.org/api/feed/audiobooks?limit=1&format=json | anonymous access works and the record carries url_zip_file pointing at archive.org/compress/<item>/ |
More fields (3)
{
"max_limit": 500,
"highest_project_id": 23017,
"audio_host": "archive.org"
}