Releases: mindsdb/mindsdb
v25.7.3.0
TL;DR
Changes to SQL
"IF NOT EXISTS" clause is now supported for "CREATE AGENT" command, enabling users to create agents without worrying about duplicate entries. We've also made a fix for an incorrect error message in "DELETE TABLE" when the database does not exist -- we're now providing the correct information.
Changes to Integrations
Improvements were made in Salesforce metadata, Snowflake Data Catalog stats error, table name specification for lower case, and Slack list pagination. To facilitate easier S3 integration, a public_url parameter has been introduced. Additionally, Youtube integration now supports search functionality. UV has also been downgraded for performance optimization.
Changes to Knowledge Bases
Changes were made to whitelist embedding providers, providing a more controlled and secure environment. New documentation has also been provided covering knowledge base model usage, SQL algebra, and examples in KB query documents.
Bug Fixes and Improvements
A bug in the vectordb handler has been fixed along with improvements to Docker and Local Installations for launching all APIs. The Llama index has been bumped from 0.12.28 to 0.12.41 in the mindsdb/integrations/handlers/llama_index_handler. In addition, structured JSON response output from LLMs has been implemented.
Changes to Documentation
Updates in documentation includes instructions about mindsdb as a fully qualified engine, refined model parameters in agent documents, added Light Light Mind (litellm) documents, and syntax for adding data to agents. A spelling mistake in integrations MDX has also been corrected.
What's Changed
- Fix inserting to vectordb handler by @ea-rus in #11267
- docs -- kb models usage by @martyna-mindsdb in #11276
- docs -- refined model params in agent docs by @martyna-mindsdb in #11272
- docs -- added litellm docs by @martyna-mindsdb in #11260
- docs -- mindsdb as fqe by @martyna-mindsdb in #11284
- docs -- added * syntax for adding data to agents by @martyna-mindsdb in #11288
- Downgrade UV by @hamishfagg in #11281
- Fix/chan notify by @hamishfagg in #11269
IF NOT EXISTS
forCREATE AGENT
by @StpMax in #11271- Updated Docker and Local Installations to Launch All APIs by @MinuraPunchihewa in #11283
- Added the public_url Parameter to the S3 Integration by @MinuraPunchihewa in #10640
- KB: whitelist of embedding providers by @ea-rus in #11275
- Fix Salesforce Metadata by @ZoranPandovski in #11274
- Fix
InterfaceError
for Snowflake Data Catalog stats by @ZoranPandovski in #11173 - Make sure table name is lowercase by @ZoranPandovski in #11273
- docs -- added sql algebra and examples to kb query docs by @martyna-mindsdb in #11286
- Correct error message for
DELETE TABLE
if database not exists by @StpMax in #11287 - Bump llama-index from 0.12.28 to 0.12.41 in /mindsdb/integrations/handlers/llama_index_handler by @dependabot[bot] in #11280
- Fixed a spelling issue in integrations.mdx by @arun-prasath2005 in #10882
- Added Structured JSON response output from LLMs by @PriyanshuPz in #11178
- Added support for search in Youtube integration by @PriyanshuPz in #11195
- Update test_on_deploy.yml by @ZoranPandovski in #11042
- Prevent template injection by @StpMax in #11279
- fix slack list pagination by @omerc7 in #10895
New Contributors
- @arun-prasath2005 made their first contribution in #10882
- @omerc7 made their first contribution in #10895
Full Changelog: v25.7.2.0...v25.7.3.0
v25.7.2.0
TL;DR
Changes to SQL
- Updated agents documentation related to SQL enhancing user understanding about connecting and utilizing various databases.
- An upgrade in error handling for internal SQL query processing which makes it more robust and user-friendly.
- Added the feature of keyword search that will allow using SQL for more dynamic and flexible data searching.
Changes to Integrations
- Support added for Google Gemini Model in TO_MARKDOWN() function. This expansion allows users to process more types of data in structured formats.
- TO_MARKDOWN() support has been extended for XML/Nessus Files, widening the variety of file types that can be exploited by this versatile function.
- Fixed an issue with Salesforce Data Catalog, restoring its compatibility and performance.
- Config update behavior has been changed to overwrite default model parameters. This offers users more control and flexibility when configuring their models.
Changes to Knowledge Bases (KB)
- An overhaul to KB error messages was done for more informative and helpful error prompts.
- Updated documentation to explain the from_sql info in the knowledge bases section, providing detailed instructions on its usage.
- Introduced unit tests for KB Evaluate function to ensure its reliable performance.
Bug Fixes and Improvements
- To ensure data integrity, preventive measures against information exposure through exception have been implemented.
- The
mcp
dependency has been updated to v1.10.1 for improved stability and optimized performance. - Refactoring of agent streaming methods using httpx and fix serialization in the A2A server was done for enhanced efficiency.
- A2A server configuration location was fixed for smoother operation.
Changes to Documentation
- Provided documentation on how to incorporate url_file_upload into config.json.
- Updated model types documentation for a more in-depth understanding of different models and their uses.
- Documentation about logging configuration was added for assistance in setting and managing log protocols.
- Knowledge base documentation was split into multiple pages for an easier and better understanding.
What's Changed
- docs -- updated docs for agents via sql by @martyna-mindsdb in #11231
- KB error messages by @ea-rus in #11198
- Enabled Google Gemini Model Support in TO_MARKDOWN() Function by @MinuraPunchihewa in #11176
- Prevent information exposure through exception by @StpMax in #11190
- docs -- url_file_upload into config.json by @martyna-mindsdb in #11244
- Changed Config Update Behavior to Overwrite Default Model Parameters by @MinuraPunchihewa in #11233
- Extended TO_MARKDOWN() with Support for XML/Nessus Files by @MinuraPunchihewa in #11228
- docs -- evaluate kb updates to from_sql info by @martyna-mindsdb in #11250
- docs -- updated model types docs by @martyna-mindsdb in #11249
- docs -- logging config by @martyna-mindsdb in #11251
- KB Evaluate unit tests by @ea-rus in #11245
- Bump
mcp
dependency to v1.10.1 by @StpMax in #11254 - Better error handling for internal sql query processing by @StpMax in #11236
- adding keyword search by @fshabashev in #11221
- Fix Salesforce Data Catalog by @ZoranPandovski in #11256
- A2A server: refactor agent streaming methods using httpx and fix serialisation by @dusvyat in #11247
- docs -- split of kb docs into pages by @martyna-mindsdb in #11258
- Fix a2a server: changed location of a2a config by @ea-rus in #11170
- Fix CI error cannot import name 'KB_TO_VECTORDB_COLUMNS' by @ea-rus in #11263
Full Changelog: v25.7.1.0...v25.7.2.0
v25.7.1.0
TL;DR
Changes to SQL
- Processing for queries containing
INTERSECT
has been added. This enables the users to find common entries from different sets of data. - Fixed error with
show columns from information_schema.{table}
. This provides users with the correct information about the table's structure and content. - Improved error message formatting for the database. This will give users a clearer understanding of any issues that arise during the database operation.
- Updated config options for file upload via URL. This change gives users more flexibility and more secure methods to upload files.
Changes to Integrations
- Fixed a bug regarding subscription in PostgreSQL. This fix will result in better and more reliable mechanisms for real-time syncing and publishing data.
- Important fixes were done on Big Query Data Catalog enhancing seamless integration with Google's BigQuery.
Changes to Knowledge Bases
- Notable improvements have been made in terms of functionality and performance of Knowledge Bases (KB). Including search and evaluate features improvements, metadata updates, filter restriction by unknown metadata keys, algebra operations, and more. Such changes are expected to improve user experience with KB in terms of search accuracy, metadata handling, and manipulation.
- The creation agent now supports wildcards in table/KB names which increases usability, efficiency, and flexibility in defining database schemas.
Bug Fixes and Improvements
- The agent can now handle wrong quoting table names more effectively. This reduces the error rate when handling table names.
- Temporary disable
huggingface_handler
. This might contribute to better resource allocation and overall performance. - Parameters in Google Gemini model examples have been updated for consistency which led to improved model performance and usability.
Changes to Documentation
- A considerable amount of updates have been provided on the documentation end. Updates include Docker installation with APIs, mindsDB variables, evaluate docs, Data Catalogue, and more. This strengthens the reference for users and helps them understand all aspects of MindsDB in better detail.
What's Changed
- Temporary disable
huggingface_handler
by @StpMax in #11155 - Added processing for queries containing
INTERSECT
by @StpMax in #11052 - fix: update parameter names in Google Gemini model examples for consistency by @PriyanshuPz in #11165
- Data Catlogue Docs by @ZoranPandovski in #11131
- Changes to
test_mysql_api.py
based on db driver features by @StpMax in #11175 - docs -- alter view by @martyna-mindsdb in #11168
- Fix KB evaluation: evaluation didnt make search in KB by @ea-rus in #11188
- Create agent: support wildcards in tables / KB names by @ea-rus in #11067
- Adding keyword search by @fshabashev in #11114
- Create agent:
data
parameter by @ea-rus in #11039 - KB Evaluate: fix when base_url is empy string by @ea-rus in #11192
- KB algebra by @ea-rus in #11132
- Fix Subscribe in PostgreSQL by @ZoranPandovski in #11150
- Big Query Data Catalog fixes by @ZoranPandovski in #11177
- KB metadata keys: service values are underscored by @ea-rus in #10939
- KB restrict filter by unknown metadata keys by @ea-rus in #11019
- KB: set created_at and updated_at in metadata by @ea-rus in #10950
- Fix
show columns from information_schema.{table}
by @StpMax in #11058 - Better db error message formatting by @StpMax in #11200
- Agent: handle wrong quoting table names by llm by @ea-rus in #11191
- New config options for file upload via URL by @StpMax in #11179
- Use parser tests with the same version as installed version by @ea-rus in #11215
- Improve MCP tools description by @StpMax in #11218
- feat: add git feature to devcontainer configuration by @parthiv11 in #10398
- update static IPs to new cluster ones by @hamishfagg in #11224
- Fix KB reranker: show error, support gemini by @ea-rus in #11229
- Create agent: llm parameters as dict by @ea-rus in #11043
- docs -- docker installation with apis by @martyna-mindsdb in #11214
- docs -- mindsdb variables by @martyna-mindsdb in #11216
- docs -- changed title of evaluate docs by @martyna-mindsdb in #11180
- docs -- kb via rest api by @martyna-mindsdb in #11189
- docs -- updated dbeaver conn page by @martyna-mindsdb in #11237
New Contributors
- @PriyanshuPz made their first contribution in #11165
Full Changelog: v25.6.4.0...v25.7.1.0
v25.6.4.0
TL;DR
Changes to SQL
- MindsDB's information schema now includes the Data Catalog, increasing transparency for the users and simplifying data tracking.
- The SQL model now supports Gemini 2.5, providing enhanced features and improved stability.
Changes to Integrations
- Ludwig handler integration received an update with an upgraded ray, which results in better compatibility and increased performance.
Changes to Knowledge Bases
- Correction of agent stripping input misuse in the kb_query_tool will help avoid unnecessary errors.
- An override of Langchain base prompts was included, adding more proficiency to the KB tools.
- The issue with evaluating KB for MySQL databases was fixed enhancing efficacy.
Bug Fixes and Improvements
- Corrections have been made to the default values and descriptions for SF columns metadata, providing more detailed data insights.
- Improvement of error messages now offers developers a clearer understanding of problems.
- Fixed handling of default model configuration parameters is providing a more smooth working process.
- Table name checking in lower case in API handler provides better string matching and avoids possible errors/casesensivity issues.
Changes to Documentation
- Documentation updates including notes on starting MindsDB, changing relevance_threshold into relevance, removing reference to minds endpoints, and many others offer clearer guidelines.
- A note was added on Mindsdb Cloud Provider (MCP) authentication. Furthermore, the knowledge base has been updated for the Snowflake provider.
- It was noted that Ollama support has been removed from knowledge bases.
What's Changed
- Exposed the Data Catalog as Part of MindsDB's Information Schema by @MinuraPunchihewa in #11087
- Bump ray from 2.8.1 to 2.43.0 in /mindsdb/integrations/handlers/ludwig_handler by @dependabot in #10543
- Bump brace-expansion from 1.1.11 to 1.1.12 in /docs by @dependabot in #11072
- Add Support for Stable Gemini 2.5 Models by @noname4life in #11117
- Fix agent with KB: strip input of kb_query_tool by @ea-rus in #11136
- Add default value and description for SF columns metadata by @ZoranPandovski in #11135
- docs -- starting mindsdb with requires apis by @martyna-mindsdb in #11130
- kb docs -- changed relevance_threshold into relevance by @martyna-mindsdb in #11129
- docs for kbs -- removed ollama support by @martyna-mindsdb in #11139
- Override Langchain Base Prompts & Only Include KB Tools if Needed by @tmichaeldb in #11123
- Fix Evaluate KB for mysql database by @ea-rus in #11137
- docs for extended config -- default reranking model by @martyna-mindsdb in #10990
- docs -- mcp auth note by @martyna-mindsdb in #11149
- docs -- removed ref to minds endpoints by @martyna-mindsdb in #11151
- docs -- kb update for snowflake provider by @martyna-mindsdb in #11152
- Better error messages by @StpMax in #11121
- Check table name in lower case in API Handler by @StpMax in #11111
- Fixed Handling of Default Model Configuration Parameters by @MinuraPunchihewa in #11161
- Release 25.6.4.0 by @MinuraPunchihewa in #11171
New Contributors
- @noname4life made their first contribution in #11117
Full Changelog: v25.6.3.1...v25.6.4.0
v25.6.3.1
TL;DR
Bug Fixes and Improvements
- An enhancement was made in handling data extraction from tables in Knowledge Bases, providing increased reliability.
- Developers can now use the Default LLM as the Default Re-ranking Model—granting them more flexibility and customizability.
- There were issues with Knowledge Base insert operations via SQL Queries and Raw Data, which have now been addressed for smoother operations.
Changes to SQL
- A key change in SQL is replacing
locals()
withconnection.register
for registering duckdb tables, which greatly improves readability and organization.
Changes to Integrations
- A few major updates in dependencies have been executed to enhance system compatibility and maintain software hygiene. Protobuf has been bumped from 3.20.3 to 4.25.8 for Lindorm handler, and setuptools have been increased from 75.8.2 to 78.1.1 in requirements.
- Additionally, requests have also been updated from 2.32.3 to 2.32.4 in requirements to keep up with changes in the software environment.
What's Changed
- Fixes for agent: getting data from tables in KBs by @ea-rus in #11115
- Bump protobuf from 3.20.3 to 4.25.8 in /mindsdb/integrations/handlers/lindorm_handler by @dependabot in #11109
- Bump requests from 2.32.3 to 2.32.4 in /requirements by @dependabot in #11038
- Added the Option to Use Default LLM as the Default Re-ranking Model by @MinuraPunchihewa in #11119
- Replace
locals()
toconnection.register
when register duckdb tables by @StpMax in #11099 - Fixed Knowledge Base Insert Operations via SQL Queries and Raw Data by @MinuraPunchihewa in #11127
- Bump setuptools from 75.8.2 to 78.1.1 in /requirements by @dependabot in #10991
Full Changelog: v25.6.3.0...v25.6.3.1
v25.6.3.0
TL;DR
Bug Fixes and Improvements
- The output of
Evaluate
has been fixed, by removingretrieved_in_top_k
, which helps to make results cleaner. - Invalid model error with Azure openai models has been ironed out to allow smoother execution.
- Successful completion is now a requirement for test setup, improving the reliability of tests.
- The deletion issue within ML engine with data catalog enabled has been resolved.
- Various fixes have been implemented, like correcting the 'SELECT' from API Resources, fixing the querying of Snowflake columns with special characters, and fixing the
ssl
option of MySQL handler. These contribute to a smoother, more reliable user experience.
Changes to SQL
- An adjustment has been made to reduce token usage in generating schema information for agents, which can lead to performance improvements.
Changes to Integrations
- Google BigQuery is now a supported source in Data Catalog, expanding the usability of the platform.
- Fixes to 'Agent Salesforce' have been introduced to enhance functionality.
- The llama_index_handler has been updated from version 0.12.21 to 0.12.28, ensuring that the handler operates using the latest available features and fixes.
Changes to Knowledge Bases
- Knowledge base documentation has gone through numerous updates, for instance, 'evaluate kb', 'evaluate syntax', and new links to AI integrations have been added. This increases the resourcefulness of the platform.
- Knowledge base for Python SDK has been hidden.
- The REST API Endpoints for Knowledge Bases has been updated for better accessibility and usability.
- The platform now provides functionality for enhanced conversation history handling in MindsDBAgent and AgentTaskManager, creating a more efficient system.
Changes to Documentation
- Documentation has been updated with usage examples in 'ollama knowledge base', creating a more user-friendly guide.
What's Changed
- Fix Evaluate output: remove retrieved_in_top_k by @ea-rus in #11050
- Reduced Token Usage in Generating Schema Information for Agents by @MinuraPunchihewa in #11049
- remove prod slack msg by @hamishfagg in #11047
- Update constants.py by @ZoranPandovski in #11045
- fix invalid model error on azure openai models by @fshabashev in #11075
- hide kb docs of python sdk by @martyna-mindsdb in #11078
- kb docs -- evaluate kb by @martyna-mindsdb in #11035
- Require success for test setup by @hamishfagg in #11051
- kb docs -- evaluate syntax by @martyna-mindsdb in #11080
- Fix ML engine deletion with data catalog enabled by @hamishfagg in #11082
- Fixed SELECT from API Resources by @MinuraPunchihewa in #11084
- Updated the REST API Endpoints for Knowledge Bases by @MinuraPunchihewa in #10813
- Enabled Google BigQuery as a Supported Source in Data Catalog by @MinuraPunchihewa in #10947
- A2A: Enhance conversation history handling in MindsDBAgent and AgentTaskManager by @setohe0909 in #11083
- Bump llama-index from 0.12.21 to 0.12.28 in /mindsdb/integrations/handlers/llama_index_handler by @dependabot in #11032
- update documentation with ollama knowledge base usage example by @fshabashev in #11103
- Agent Salesforce fixes by @ZoranPandovski in #11077
- Fixed Querying Snowflake Columns with Special Characters by @MinuraPunchihewa in #11101
- Fix
ssl
option of mysql handler by @StpMax in #11102 - kb docs -- added links to ai integrations by @martyna-mindsdb in #11079
Full Changelog: v25.6.2.0...v25.6.3.0
v25.6.2.0
TL;DR
Changes to SQL
- Fixed Metadata Queries for Snowflake in Data Catalog, providing users with more accurate and reliable data query capabilities.
- Add support to agent for default LLM parameters from config introduces additional functionalities to the agent system making it more flexible.
- Added Support for ALTER VIEW and Update DATABASE commands in SQL to further enrich the SQL operations.
- Correct JSON types pass through MindsDB, optimizing the JSON types handling in the system.
- Improved the render of Cast functional select targets for BigQuery, providing seamless experience when interacting with BigQuery.
Changes to Integrations
- Enable Data Catalog Support for the New Agent Implementation improves the integral functionalities between Data Catalog and the new agent system.
- Salesforce has been enabled as a Supported Source in Data Catalog, offering users additional options in catalog data sources.
- Added more OpenAI models for agents improves the AI model support for agents.
- Agents Data Catalog and Data Catalog V1 have been added, which could enhance the functionality and user experiences in the data catalog area.
Changes to Knowledge Bases
- Evaluate the knowledge base and fixing types in knowledge base help to optimize the performance of knowledge base.
- Add empty reranking for knowledge base (KB) controller as well as adding parameters to the select query to turn off reranking provide users with more customizable options.
- Multiple updates on KB Docs improve the documentation on knowledge base.
Bug Fixes and Improvements
- Fixing connect from Pycharm to MySQL server would make the connection from Pycharm to MySQL server smoother.
- Add API Key handling in agents, fixes for file logging in Windows, and added checks for not-empty and unique file names when upload a file have enhanced the security and reliability of the system.
- Show exceptions of LLM function and set progress for insert without partitions help to improve debugging and monitoring.
- The infrastructure has removed prod deploy while enables auto migrations in cloud to optimize the service deployment process.
- Add /status endpoint for A2A to provide server health information enriches the health check functions for the server status.
- Various testing fixes and simplify testing can prevent potential bugs and improve the performance of the system.
Changes to Documentation
- Enabled several document updates including creating index on knowledge base, metadata, provider name fix and supported models, which provides clearer guidelines and detailed instructions to users.
What's Changed
- Data Catalog V1 by @MinuraPunchihewa in #10862
.....................
..................... - Support of
vector
type by @StpMax in #11030
Full Changelog: v25.5.4.2...v25.6.2.0
v25.5.4.2
TL;DR
Bug Fixes and Improvements
The logic of acquiring the A2A configuration has been enhanced. The recent update will help to streamline the experience for users working with configuration files. The details of this modification can be found in the following "What's Changed" section.
What's Changed
- fix getting a2a config by @hamishfagg in #10942
Full Changelog: v25.5.4.1...v25.5.4.2
v25.5.4.1
TL;DR
Changes to Integrations
Three new integration updates have been introduced:
- Added Writer support to agents for better data management.
- An improved configuration and Docker setup were added with A2A API support.
- MindsDB now extends to the cloud with the addition of SalesForce.
Changes to Knowledge Bases
A few adjustments have been made to improve knowledge bases:
- A blueprint for accessing llm providers using a litellm handler has been implemented.
- The default Re-ranking Model was enabled, enhancing the quality of information retrieval.
- The utilization of MindsDB models for Knowledge Bases is now disabled, ensuring a more streamlined process.
Bug Fixes and Improvements
Several bugs have been fixed and improvements were added:
- There has been a fix for a previously existing 'insert bug'.
- A change was made so that Slack does not receive messages for cancelled workflow runs, reducing unnecessary notifications.
- An A2A client was removed due to its lack of use, optimizing resources.
Changes to Documentation
Two updates were made to the documentation:
- There's a new set of documents detailing the 'mcp'.
- The documentation for agents has also been improved.
Changes to SQL
No SQL changes were made in this update.
What's Changed
- mcp docs by @martyna-mindsdb in #10876
- docs for agents by @martyna-mindsdb in #10902
- Run whole integration test folder by @hamishfagg in #10904
- Use ruff for linting and formatting by @hamishfagg in #10916
- Don't send slack messages for cancelled workflow runs by @hamishfagg in #10918
- Use litellm handler in KB for accessing to llm providers by @ea-rus in #10855
- Enabled Default Re-ranking Model for Knowledge Bases by @MinuraPunchihewa in #10848
- add ruff dep and make a format pre-commit hook by @hamishfagg in #10925
- Add A2A API support to configuration and Docker setup by @dusvyat in #10919
- remove a2a client as it isn't being used by @dusvyat in #10927
- Add Writer support to agents by @ZoranPandovski in #10915
- fixing insert bug by @fshabashev in #10931
- Disabled Use of MindsDB Models for Knowledge Bases by @MinuraPunchihewa in #10923
- Add pyarrow to main requirements by @ZoranPandovski in #10924
- Add SalesForce to cloud by @ZoranPandovski in #10938
Full Changelog: v25.5.4.0...v25.5.4.1
v25.5.4.0
TL;DR:
Bug Fixes and Improvements
- Fixed issues on release testing, enhancing the process of releasing newer versions.
- Enhanced staging test failure messages and added new release messages, leading to a more smooth and informed development process.
- Improved file buffer closing during file upload, adding to the reliability of file operations.
- Handled the Transformer to be optional. This provides more flexibility when utilizing transformers in our projects.
- Improved stream fetching of data from the data handler, providing a more efficient data handling solution.
- Fixed import issues of Handler Utilities, leading to a smoother start of your projects.
- Fixed bugs and perfections related to the integration of Handler Utilities.
Changes to SQL
- Provided support variables in SQL, making SQL execution more dynamic and flexible.
- Further improvements have been made to our SQL version, aligning it with industry standards and modern SQL features.
- Allowed
INSERT
without a column list, providing more flexibility in SQL operations. - Register inserts from selects in queries table integrating, this helps in tracing and optimizing DB operations.
- Introduced 'Create index' altering the table query to set the table's dimension for index creation.
Changes to Integrations
- Added an A2A server. This will provide better interactive and collaborative capabilities for the MindsDB ecosystem.
- Implemented token-based auth for the MCP server. This improves security features and safeguards user data more effectively.
Changes to Knowledge Bases
- Updated the knowledge base docs example, making it easier for users to understand and use MindsDB's knowledge base.
- Added JSON chunking support for the knowledge base, which will allows for more efficient storage and retrieval of data.
Changes to Documentation
- Conducted a major refactoring of the MindsDB docs. The documentation is now much more navigable and user-friendly.
- Updated the mindsdb and minds docs, ensuring they provide accurate and updated information.
- Included many updates to the documentation files and also included new copies to specific doc pages.
What's Changed
- Fix testing on release by @hamishfagg in #10842
- Fix staging test fail msg, and add release msg by @hamishfagg in #10843
- Support variables in sql by @ea-rus in #10829
- Ensure file buffer is flushed before closing in file upload by @ZoranPandovski in #10845
- Transformers optional by @ZoranPandovski in #10847
- Stream fetching data from data handler by @ea-rus in #10724
- Set uv in deploy action by @ZoranPandovski in #10846
- kb docs example update by @martyna-mindsdb in #10853
- docs for agents and chatbots updates by @martyna-mindsdb in #10816
- docs refactor to merge mindsdb and minds docs by @martyna-mindsdb in #10783
- add codeowners by @hamishfagg in #10850
- Fixed Imports of Handler Utilities by @MinuraPunchihewa in #10856
- Add transformers to main requirements by @ZoranPandovski in #10859
- Register inserts from selects in queries table by @ea-rus in #10840
- Adding create index by @fshabashev in #10854
- added copy to doc pages by @martyna-mindsdb in #10868
- docs into connect/unify/respond sections by @martyna-mindsdb in #10865
- Add 'checkpoint' migration by @StpMax in #10863
- docs files updates by @martyna-mindsdb in #10871
- docs langchain updates by @martyna-mindsdb in #10870
- temp hide agents docs by @martyna-mindsdb in #10877
- Update README.md by @torrmal in #10849
- A2A server by @dusvyat in #10857
- Allow
INSERT
without columns list by @StpMax in #10872 - bugfix: adding alter table query to set table dimension for index cre… by @fshabashev in #10886
- Update sample-database.mdx by @martyna-mindsdb in #10901
- Add JSON chunking support for KB by @dusvyat in #10823
- Testing cleanup by @hamishfagg in #10793
- Remove flake8 exceptions and fix issues by @hamishfagg in #10896
- Token based auth for MCP server by @StpMax in #10874
- Improvements of column type mapping in the MySQL API. by @StpMax in #10716
- Add back A2A unit tests and remove outdated A2A HTTP endpoint tests. by @dusvyat in #10909
- Update SQL version by @ZoranPandovski in #10910
Full Changelog: v25.5.3.0...v25.5.4.0