Skip to main content

v1.0.202

Date released: November 12, 2025

New features and enhancements​

Agents​

MongoDB agent​

The MongoDB Agent is an AI-powered system designed to answer natural language queries by interacting directly with MongoDB databases.

Core functionality​

  • Schema collection: It gathers schema information from designated MongoDB collections using the Datagol API or by directly inspecting the MongoDB instance.

  • Query generation: The system uses Large Language Models (LLMs) to identify schema gaps and then generate optimized Python data collection scripts.

  • Efficient data retrieval: These scripts execute MongoDB queries using flexible string matching, filters, and complex aggregation pipelines for high efficiency.

  • Data processing: Retrieved data is loaded into pandas DataFrames for subsequent LLM-based analysis.

Output and experience​

  • Results presentation: The final results are returned as formatted tables accompanied by relevant metadata.

  • Visual cards: For smaller datasets (fewer than 10 rows), the agent automatically generates visual cards to enhance readability.

  • Advanced features: The agent supports schema caching, handles cross-collection joins, and provides a superior user experience through streaming responses with progress updates throughout the entire workflow.

Web Connector​

The Web Connector has been upgraded to enhance its ability to extract rich content from web URLs by integrating the unstructured library. This enhancement ensures that web data sources provide more comprehensive and structured inputs for subsequent analysis, including both extracted text and image descriptions.

Was this helpful?