The Machine Learning Recipe

How do we get our Spicy Estimates?

The SPICYEST machine learning model receives and prepares various sources of data. Then, it mixes data from traits, listings, and historical sales to create our Spicy Estimates.


collections processed


total estimates




estimate refresh

Accurate Prices

Because the ingredients that make up the Spicy Estimates are the best that can be gathered, SPICYEST tends to have a median relative error (MRE) of just 5-10%. This means estimates are usually within 5-10% of the next sale price.

Live Data

The Spicy Estimates are constantly being processed real-time to guarantee users the freshest appraisals. When new transactions happen within a collection, trust that the SPICYEST model has already been retrained with this new data.

What else do we serve?

Price Breakdown

Understand why your estimate looks the way it does; the breakdown shows how the SPICYEST model believes each trait contributed to the final price.

Price Range

Feel confident with our price estimate by knowing the highest and lowest value that the NFT is possibly worth.


Explore other similar NFTs in the collection based on their own traits and estimates.

Get real-time NFT price estimates

Check out our public pricing oracle app and get access to over 25 million real-time estimates.