MindsDB Release Notes - August 2025

August updates to the MindsDB Open-Source AI platform are here—usability improvements, new integrations, and smarter SQL workflows. Explore everything in the current release, v25.8.2.0, below.

Knowledge Bases

The latest update brings several improvements to MindsDB’s Knowledge Bases, making them easier to use and understand. Documentation has been refreshed with a clearer introduction and expanded guidance on hybrid search, helping users get up to speed faster. On the technical side, error handling has been improved—queries made against non-existent databases now return more informative error messages, giving users better feedback and guidance when something goes wrong. These changes make the Knowledge Base experience both more accessible and more reliable.

Integrations

As part of this release, MindsDB continues to refine its integrations and expand database support. The AnyScale handler has been removed following an internal review of supported connectors, streamlining the platform and ensuring long-term stability. At the same time, support for IBM Db2 databases has been added, giving users more flexibility when working across diverse enterprise environments. Additionally, the Hugging Face handler has been reintroduced after key bug fixes, restoring its functionality and improving the overall experience for users who rely on Hugging Face models within MindsDB. These changes reflect a balance of pruning, extending, and enhancing—ensuring that MindsDB evolves in ways that keep it both versatile and reliable.

Bug Fixes and Improvements

A series of important bug fixes and UI improvements aimed at making MindsDB more stable and user-friendly have been made. A bug in the Hugging Face handler that triggered import errors has been resolved, restoring reliability for model integrations. The Docker build process was updated to improve the graphical interface, though later adjustments reverted those changes to maintain overall stability. On the API side, creating VIEWS via REST-API now includes SQL validation, preventing errors before they occur. Finally, the calculation of processed rows has been refined, ensuring more accurate reporting for users. Collectively, these fixes tighten up the user experience across multiple touch points, from model handling to database operations.

Documentation

The documentation improvements designed  aims to make MindsDB easier to learn and work with. Introductory sections have been refreshed for clarity, while pages like the LangChain integration have been fixed and expanded. Updates now cover new query parameters for agents, support for OpenAI-compatible APIs via the OpenAI handler, and detailed guidance for working with Snowflake both as a data source and in Knowledge Bases. Additional improvements include clearer instructions for using the REST API for login, new examples for Knowledge Base docs via the Python SDK, and an updated page for building custom database handlers. Together, these changes make the docs more comprehensive and approachable, giving users a smoother path from getting started to building advanced integrations.

SQL Operations

The SQL prompt has been enhanced to make queries more effective and reliable, giving users better results when working with Salesforce and related data. Issues with Common Table Expressions (CTEs) have been resolved, ensuring smoother execution of complex queries that rely on nested logic. Finally, MySQL functions are now mapped to DuckDB, making it easier for users to run familiar MySQL-style queries while taking advantage of DuckDB’s performance and compatibility. These improvements collectively reduce friction and expand flexibility for anyone working across multiple SQL dialects within MindsDB.

OS GUI

This release also delivers several improvements to MindsDB’s GUI, making it easier and more intuitive to use. A long-standing bug that caused the Respond tab to fail when running over SSL has been fixed, so SSL is now fully supported. We’ve also expanded SSL connection support to make MindsDB more flexible when accessing the GUI. You can now connect using combinations like domain + port, IP + port, or simply domain on its own. These strategies provide more options for securely reaching the interface, giving teams greater flexibility depending on how their environment is set up. This improvement makes it easier to run MindsDB in different deployment scenarios without hitting connection roadblocks.

The Respond tab now includes a code snippet section, helping users quickly access and experiment with MindsDB's Python SDK.

For those deploying on AWS, the AMI now redirects to onboarding for a smoother setup experience. We also fixed a filtering issue in the Manage Integrations tab, ensuring that once the list of datasources is filtered, it correctly resets to the original view. Finally, the Thoughts column outputs in the Respond tab have been improved, providing clearer, more understandable feedback when interacting with the system.

We’ve also streamlined the process of adding Knowledge Bases: instead of adding an LLM directly, the “Add” button now loads the proper SQL syntax to simplify creating Knowledge Bases seamlessly.

With this release, MindsDB continues to get faster, smarter, and easier to use—whether you’re working with SQL, knowledge bases, integrations, or the UI. Each improvement, from better error handling to expanded database support, is designed to help you spend less time managing infrastructure and more time building with your data. Read the full release notes for all recent updates.

As always, we’d love to hear your feedback. Try out the new features, explore the docs, and let us know how MindsDB can make your workflow even smoother. Upgrade today and take these new capabilities for a spin—we can’t wait to see what you build.