Alchemy is a decentralized data model training platform based on on-chain secure encryption operations and random forest training models. Blockchain ecological nodes return beacons by transmitting transaction decision data to realize data retrieval, use, and confirmation. At the same time, Alchemy uses the decentralized decision-making model of the random forest deep learning algorithm to build a unique algorithm training system, and uses the beacons fed back from the data samples to correct the fitting model, thus opening a new era of decentralized data model training platforms.