MindsDB is an open-source AI query engine that lets you bring machine learning and LLM capabilities directly to your data using SQL. Founded in 2017 in Berkeley, California, the project has accumulated over 26,000 GitHub stars and positions itself as the bridge between traditional databases and modern AI. The core idea is radical in its simplicity: instead of building complex ML pipelines that extract, transform, and load data into separate systems, you query your AI models the same way you query your database — with SQL.
The federated query engine is the foundation of everything MindsDB does. It connects to over 200 data sources — PostgreSQL, MySQL, MongoDB, Salesforce, Shopify, Google Sheets, Slack, and dozens more — and lets you query them all using a single SQL dialect. You can join data from MongoDB and Salesforce in one query without any ETL pipeline. For AI agents, this means faster response times, better accuracy, and lower token consumption because the data is accessed directly rather than being replicated into a separate system. Create views across data sources, and your agents have live access to the latest data.
MindsDB's Knowledge Base unifies structured and unstructured data, making sense of everything from database records to documents, support tickets, and Google Drive files. The Minds Cognition engine understands questions, plans data retrieval, and finds the most relevant information to respond — with full transparency into its reasoning and actions visible to operators. This is not just a database with AI bolted on; it is a query engine designed from the ground up to make AI a first-class citizen alongside traditional data access patterns.
Agent creation follows the same SQL-native philosophy. You define an agent with a CREATE AGENT statement specifying the LLM provider, model, API key, data sources (knowledge bases and tables), and a prompt template. The agent then has federated access to all connected data sources and can answer questions grounded in your actual business data. MindsDB supports Model Context Protocol for connecting agents with external tools, and includes workflow automation through jobs and triggers that can schedule and chain operations.
The platform handles both traditional ML and modern LLM use cases. On the ML side, MindsDB can perform predictive analytics, forecasting, anomaly detection, and multivariate time-series analysis directly within your database. On the LLM side, it integrates with OpenAI, Anthropic, and other providers to enable natural language querying, text generation, classification, and semantic search. The ability to combine both capabilities — running a prediction model against structured data and then using an LLM to explain the results in natural language — is unique.
Deployment flexibility spans the full range. The open-source version runs anywhere — on-premises, in a VPC, or serverless. The Minds Enterprise offering adds cognitive engine capabilities, advanced security and governance including GDPR and HIPAA compliance, custom user roles, personalized data permissions, and enterprise SLAs. Pricing for the enterprise tier is custom, while the open-source version is completely free. This dual-track approach lets teams start with the open-source version and upgrade to enterprise features as their needs grow.