Cropster to release AI-based first crack prediciton

Cropster has released its Artificial Intelligence (AI)-based First Crack prediction program, the second feature based on its work with machine learning. The roasting software specialist says there is more to come.

“We roasted a lot on all our three roasters last week using the latest version of the first crack prediction. And to be honest, it is scary how accurate it is now,” says beta tester Johannes Otto of Rösttrommel.

“First crack happened within two to three seconds to the prediction. And it was accurate on all our machines. Even on our small Giesen Sample roaster with a completely unknown coffee.”

In 2020, Cropster introduced an industry first – AI based curve prediction. Cropster AI solutions work with any coffee roasting machine that connects to Cropster and allows roasters to see the temperature curve up to two minutes into the future with less than 2 per cent variance.

Roasters can ‘roast ahead of the curve’ and adjust their gas/temp based on where things inside the roast machine are going versus where they are now.

With First crack Prediction, Cropster adds another leap forward. This AI based feature takes the guesswork out of the first crack to help roasters get ready well before it happens with incredible accuracy.

Roasters using Cropster are now much closer to knowing the “exact” time of first crack so they can plan and roast accordingly.

“The prediction was super beneficial when dealing with abnormal batch sizes (i.e. bigger or smaller than standard size), and when creating new roast profiles. It’s a super cool feature,” says beta tester Katie Campbell of Mission Coffee.

Cropster says first crack prediction gets roasters a peak result faster and with fewer batches, saving them time and money. Alongside roast curve prediction, it also makes dialling in a new roasting machine easier and less expensive. It also makes training staff and focusing on key aspects of the roasting process is easier.

“I like the idea, similar to curve prediction, that this feature keeps future events front of mind for the roaster. Reminding us that we are working towards time and temperature goals,” says beta tester Paul Arnephy of  Cafe Lomi.

© All Rights Reserved. BeanScene is a registered trademark of Prime Creative Media.