As Discogs buyers, we tend to prefer sellers who have multiple items from our wantlist. This allows us to order several records while minimizing shipping costs.
The Discogs engine matches a seller’s inventory to our wantlist strictly by release. If we forget to include all versions of a specific master release (such as different pressings or editions of the same album) in our wantlist, they won’t be matched with the seller’s inventory. For example, if an album has two pressings, X and Y, and you only have X on your wantlist, a seller offering Y won’t be suggested to you, even though it’s the same band and album.
The code enhances matching by using the API to fetch master IDs from your wantlist, retrieving the seller’s inventory, gathering all release versions of matched master releases, and cross-referencing them to identify matches between the wantlist and the seller’s inventory.
With Discogs Wantlist Search

With Enhanced Wantlist Search



- Clone the repository
- Log in to Discogs and generate your API Token. For more API Discogs documentation, read here
- Input YOUR_API_TOKEN and YOUR_SELLER_USERNAME(name of a shop you want to fetch data from) into wantlist_vs_seller_matcher.ipynb and execute the Jupyter Notebook!