On a mission to democratize AI.

In 2018, in Berkeley, California, Jorge Torres and Adam Carrigan released the first version of MindsDB into the world. Their starting point was an idea borrowed from Iain M. Banks' Culture novels: that AI, at its best, is a partner — a "Mind" that augments human judgment instead of replacing it.

Translating that idea into something useful for the world's data teams meant rejecting the dominant pattern of the time. The standard ML pipeline was a relay race: extract data from one system, clean it, model it somewhere else, then thread the predictions back into production. It was slow. It was fragile. It excluded the majority of teams who had the data and the questions but not the specialists to run the relay.

MindsDB took the inverse position: don't ask the data to come to the AI — bring the AI to the data. Inside PostgreSQL, MySQL, MongoDB, Snowflake, and the rest of the tools teams already lived in. Familiar interfaces, no exports, no separate stack.

What began as an open-source project grew into the most widely adopted open platform for AI analytics and data workflow automation. As the work expanded from in-database ML to LLMs to open-source AI agents, the product surface broadened with it — culminating in the rebrand to MindsHub. See how the product evolved →