NASA’s new dark energy space telescope can also detect killer asteroids
technologyreview.com·1h ago
Data Science
Simon Willison's Blog·May 21·5 min read
Datasette Agent
TL;DR
Datasette, a tool for exploring databases through a web interface, lacked conversational AI capabilities for querying data naturally. Datasette Agent combines the LLM Python library (which wraps language models) with Datasette to create an extensible AI assistant that accepts natural language questions.
✦ Why It Matters
Engineers can now build conversational database interfaces without writing custom SQL query logic or complex integrations.
Key Takeaways
- Use Datasette Agent to query your data conversationally.
- Integrate datasette-agent-charts for visual data representation.
- Experiment with building your own plugins for Datasette Agent.
- Join the #datasette-agent Discord channel for community support.
Full Summary
Datasette Agent combines Simon Willison's LLM Python library with Datasette, providing a conversational interface for querying data stored in Datasette. Users can ask questions about their data, and the system generates SQL queries to retrieve relevant information.
For example, a query about recent pelican sightings returned specific details from a blog database. The agent is extensible with plugins, including datasette-agent-charts for data visualization, and others for image generation and code execution.
The underlying technology runs on Gemini 3.1 Flash-Lite, which efficiently handles SQLite queries. Future developments include additional plugins and integration with Datasette Cloud, expanding its capabilities for users.