Torrents-CSV
torrents-csv.com · Debrid · 1 record · JSON
Torrents-CSV Debrid
caveatanonymousno cardDrequest-rate
| Free tier | No key: /service/search?q=debian&size=2 returned JSON with infohash, name, size_bytes, seeders, leechers and completed. The HTML homepage is behind an Anubis proof-of-work challenge. unit requests; metered on request-count |
|---|---|
| Price | USD 0 |
| Per-object cap | not applicable: index rows |
| Retention | rows carry a scraped_date and are kept after the swarm dies; seeders is the liveness signal |
| Egress | no byte allowance: JSON only; whether the project also publishes a bulk dump was not checked on the pages read free_allowance: none published |
| Identity | anonymous No account for either surface. The HTML surface costs a proof-of-work computation, not an identity. |
| Limits |
Anubis is a bot-defence layer, so the wall is computational rather than a quota: a browser solves it in seconds, a plain HTTP client does not. |
| Binding | request-rate The HTML surface is the thing that breaks first, and it breaks on the challenge rather than on a rate; the JSON endpoint stayed open during the same session. |
| Interface | rest /service/search?q=<query>&size=<n> returns JSON. The web front end is served behind Anubis v1.28.1, a proof-of-work challenge, which the API endpoint did not present to us. |
| Tooling | One GET against /service/search. A bulk mirror would remove the query loop entirely, but the dump's existence and size were not verified today. |
| Composes with | Search by title for a quick check, or take the bulk dump and filter locally; either way the infohash is what feeds a debrid cache check. |
| Reliability | Anubis rejects clients that disable JavaScript features (the page names JShelter), and the front end and the service endpoint have different gates, so a pipeline tested on one can fail on the other. |
| Notes | Useful because the search endpoint is one GET with a stable schema, and the caveat is that the same host gates its HTML behind a proof of work, so an unannounced change would break a scraper without warning. The bulk-dump path that TorrentClaw's docs imply was not verified here. |
| Evidence rung | documented Both surfaces were fetched on 2026-10-23: the JSON answered, the HTML returned the Anubis interstitial. A recorded run against the bulk dump would settle the dataset size. |
| Checked | 2026-10-24 (live); stale on 2026-11-23 |
| Parsed numbers |
|
Evidence (2)
| date | source | what it proved |
|---|---|---|
| 2026-10-23 | torrents-csv.com/service/search?q=debian&size=2 | HTTP 200 application/json with infohash, name, size_bytes, seeders, leechers, completed and scraped_date |
| 2026-10-23 | torrents-csv.com/ | 'Making sure you're not a bot!', Anubis version v1.28.1, proof-of-work challenge instead of the index |
More fields (3)
{
"anubis_version": "v1.28.1",
"bulk_dump": "unverified",
"api_open": true
}