Returns up to 5 recommended vinyl releases for a given Discogs release. Recommendations are scored using a three-signal algorithm: similar artists from Last.fm, same-genre releases from Discogs, and (when authenticated) a boost based on the user’s own browsing history. Authentication is optional. Providing a valid JWT personalizes the results by boosting releases in genres the user has previously browsed. Without a token, recommendations are based purely on artist similarity and genre matching. Results are cached in Redis for 1 hour. The cache key is scoped per release and per user:Documentation Index
Fetch the complete documentation index at: https://mintlify.com/akibanks/api-tienda-vinilos/llms.txt
Use this file to discover all available pages before exploring further.
- Unauthenticated:
recomendaciones:{id}:anonimo - Authenticated:
recomendaciones:{id}:{userId}
Because authenticated and unauthenticated results use separate cache keys,
logging in or out will cause the next request to compute a fresh result rather
than returning the shared anonymous cache.
Endpoint
| Property | Value |
|---|---|
| Method | GET |
| Path | /disco/:id/recomendaciones |
| Auth | Optional — Bearer JWT |
| Cache TTL | 1 hour per (release ID, user) |
Request Parameters
The Discogs release ID of the album you want recommendations for.
Example:
9579406.Authentication
Pass a JWT obtained fromPOST /login in the Authorization header to receive personalized recommendations:
Response Fields
The response is a JSON array of up to 5 release objects, sorted from highest to lowest recommendation score. Each object has the same shape as a result fromGET /buscar.
A recommended release object.
Recommendation Algorithm
How recommendations are scored and ranked
How recommendations are scored and ranked
Each candidate release accumulates a score (
peso) starting at 0. After all signals are computed, all candidates are sorted by score descending, and the top 5 are returned. The source release itself is always excluded from results.Signal 1 — Last.fm similar artists (+1 per match)The Last.fm artist.getSimilar method is called with the artist of the source release, requesting up to 5 similar artists. For each similar artist, Discogs is searched for up to 3 of their releases (per_page=3). Every release found this way earns +1 point.Signal 2 — Discogs genre and style matching (+1 per match)Discogs is searched for releases sharing the same genre (and style, if available) as the source release, sorted by have descending (per_page=5). Every release found this way earns +1 point.Because a release can appear in both signals, the maximum unauthenticated score for any single candidate is +2.Signal 3 — Authenticated user history boost (+1 per genre match)If a valid JWT is provided, the server queries the user’s browsing history (up to the 20 most recent entries, ordered by visto_en descending). For every candidate release whose genero matches any genre in that history, the candidate earns an additional +1 point.This means authenticated users can see a maximum score of +3 per candidate, and results will skew toward genres they have already explored.Final sort and sliceAll scored candidates are sorted by peso descending. Ties are broken by insertion order (which signal found the release first). The top 5 are returned.Example Requests
Anonymous request:Example Response
Error Responses
| Status | Error message | Cause |
|---|---|---|
500 | "Error al obtener recomendaciones." | Network failure, upstream API error, or unexpected exception. |