live wire
▸Snowflake agent evals can pin judge versions; unpinned metrics may roll forward and shift scoresSnowflake▸ThoughtSpot makes Spotter 3 default for new instances; it requires download rights and always shares query data with an LLMThoughtSpot documentation▸MotherDuck adds image classification and object location to prompt_jev(), returning typed SQL outputs with confidence scoresMotherDuck▸BigQuery conversational analytics adds Montréal and Toronto; mixed-region agents still default to globalGoogle Cloud Documentation▸Fabric enables Azure OpenAI Copilot access by default, but OpenAI-subprocessor and cross-region processing/storage switches remain offMicrosoft Learn▸Snowflake Cortex Agents now continue by default when named tools are inaccessible, emitting warnings instead of failing the runSnowflake documentation▸Databricks Genie One can use group memberships to tailor answers, while Unity Catalog remains the access-control boundaryDatabricks documentation▸Fabric data agents can query Power BI models with Read alone; Build and workspace access are unnecessaryMicrosoft Learn▸Fabric data-agent Git configs expose column flags, but selecting a table still includes every columnMicrosoft Learn▸Google Data Agent Kit offers remote Spark kernels, but notebook exports can still move results into local Pandas or ArrowGoogle Cloud Documentation▸Cube says production analytics agents need explicit partial-result states, caller-filtered model search and durable artifactsCube▸Thinking Machines names Bridgewater behind ReViSQL-K2.6, citing 92.97% SC-16 accuracy at 15% of frontier-model costThinking Machines Lab▸AWS says a mismatched data/tool/prompt release cut its Field Sales agent pass rate from 78% to 38% in pre-productionAWS Artificial Intelligence Blog▸NL2SQL Multi-Agent ships evaluation artifacts across BIRD, CSpider, Archer and KDD DataSpace—but no software licenseGitHub▸Snowflake agent evals can pin judge versions; unpinned metrics may roll forward and shift scoresSnowflake▸ThoughtSpot makes Spotter 3 default for new instances; it requires download rights and always shares query data with an LLMThoughtSpot documentation▸MotherDuck adds image classification and object location to prompt_jev(), returning typed SQL outputs with confidence scoresMotherDuck▸BigQuery conversational analytics adds Montréal and Toronto; mixed-region agents still default to globalGoogle Cloud Documentation▸Fabric enables Azure OpenAI Copilot access by default, but OpenAI-subprocessor and cross-region processing/storage switches remain offMicrosoft Learn▸Snowflake Cortex Agents now continue by default when named tools are inaccessible, emitting warnings instead of failing the runSnowflake documentation▸Databricks Genie One can use group memberships to tailor answers, while Unity Catalog remains the access-control boundaryDatabricks documentation▸Fabric data agents can query Power BI models with Read alone; Build and workspace access are unnecessaryMicrosoft Learn▸Fabric data-agent Git configs expose column flags, but selecting a table still includes every columnMicrosoft Learn▸Google Data Agent Kit offers remote Spark kernels, but notebook exports can still move results into local Pandas or ArrowGoogle Cloud Documentation▸Cube says production analytics agents need explicit partial-result states, caller-filtered model search and durable artifactsCube▸Thinking Machines names Bridgewater behind ReViSQL-K2.6, citing 92.97% SC-16 accuracy at 15% of frontier-model costThinking Machines Lab▸AWS says a mismatched data/tool/prompt release cut its Field Sales agent pass rate from 78% to 38% in pre-productionAWS Artificial Intelligence Blog▸NL2SQL Multi-Agent ships evaluation artifacts across BIRD, CSpider, Archer and KDD DataSpace—but no software licenseGitHub
nl2sql.ai
newsDirectory alert

Vanna’s repository is archived while its product surfaces still look active

The 23,800-star text-to-SQL project now carries GitHub’s read-only flag, but its docs and package metadata do not yet explain the maintenance change.

Archived repo beside active-looking documentation and package metadata.
Side by side: what changed
By The Tools Desk· Aug 23, 2026the quick take — two AI hosts go live when you do

The main Vanna repository is archived. GitHub’s repository record marks it read-only, shows 23,821 stars and 2,485 forks, and dates the last push to February 2, 2026. That is a material maintenance signal for one of the most visible open-source text-to-SQL projects—even though the rest of Vanna’s public surface still reads like a current product.

The documentation remains online and introduces Vanna 2.0 as a user-aware AI agent framework that connects language models to databases. It describes tool execution, learned tool memory and permissions that flow through the system. The PyPI package page likewise presents Vanna 2.0 as an installable framework for natural-language-to-SQL, user-aware permissions, streaming tables and charts, and FastAPI or Flask integration.

GitHub’s release history puts the latest tagged version at v2.0.2, published on February 2. Its only listed change is a fix for unnecessary embedding-function initialization during ChromaDB collection retrieval. The repository still carries an MIT license, so archiving does not prevent users from forking or maintaining the code themselves. It does, however, prevent new issues, pull requests and repository changes unless the owner reverses the archive setting.

What users should do now

Do not interpret live documentation as evidence of active maintenance. Existing deployments should pin Vanna and its transitive dependencies, inventory open security and compatibility risks, and test a clean installation from the exact lockfile used in production. Teams evaluating Vanna should ask the maintainer whether development has moved to another repository or product before making the framework a long-lived dependency.

The project’s own migration notes make that diligence more important. Vanna 2.0 is described as a complete rewrite from the 0.x VannaBase API to an agent-based architecture, with a legacy adapter offered as a gradual bridge. A codebase that has already crossed a major compatibility boundary needs an explicit maintenance path; an archived repository with active-looking docs leaves that path unclear.

The nl2sql.ai directory now marks Vanna’s repository as archived while preserving its MIT license, current documentation link and latest verified release. That distinction matters: the software remains available, but availability and maintenance are not the same thing.

sources

Filed by The Tools Desk. Corrections: desk@nl2sql.ai · Our standards →

comments · 0

    Comments are moderated before they appear. Your email is used once to confirm it is you — never shown, never sold. Corrections and questions get an answer from the desk when we have one.

    more from the desk