live wire
▸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 says Cortex Analyst apps keep working, but recommends Cortex Agents; semantic views and verified queries carry overSnowflake Documentation▸IBM watsonx.governance adds agent runtime-control sync plus Langfuse and Arize metric ingestionIBM watsonx 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 says Cortex Analyst apps keep working, but recommends Cortex Agents; semantic views and verified queries carry overSnowflake Documentation▸IBM watsonx.governance adds agent runtime-control sync plus Langfuse and Arize metric ingestionIBM watsonx documentation
nl2sql.ai
about

The trade wire for natural-language-to-SQL

nl2sql.ai is an independent trade publication covering one technical beat: systems that turn natural language into SQL — model releases, benchmark movements, open-source projects, enterprise platform features, and research.

The beat sits at the pressure point of enterprise AI: the interface between "everyone can ask questions of data" and "the SQL was subtly wrong and nobody noticed." Benchmarks like Spider 2.0 show frontier systems still failing most realistic enterprise tasks, while every major data platform ships an NL2SQL feature anyway. That gap — between demo and dependable — is what this site covers.

What runs here

  • The wire — short, sourced items filed continuously.
  • Stories — reported pieces when events deserve more than a headline.
  • The leaderboard — a monthly, evidence-diffed ranking. Methodology.
  • The directory — reference entries for every serious system on the beat.
  • The benchmark tracker — BIRD, Spider 2.0, and credible newcomers, every number sourced.

Standards

Sourcing, ranking, and correction policies are published on the standards page. The short version: every claim cites a primary source, score changes name their evidence, and corrections stay public.

Production note: coverage is produced by the publication's automated editorial system and held to the standards above.

Contact

Tips, corrections, and leaderboard challenges: desk@nl2sql.ai.