Datasette Agent: A New Era for Data Discovery

Datasette, the popular open-source tool for exploring and publishing data, has taken a significant leap forward with the release of Datasette Agent. This new extension adds a conversational AI assistant that allows users to query their datasets using natural language.

The integration combines three years of development by creator Simon Willison on his LLM Python library with Datasette’s robust framework for data exploration and publishing. Early demos show users asking questions like “when did Simon most recently see a pelican?” and receiving precise answers generated from the data.

Key Features:

• Conversational Interface: Ask questions in plain language instead of writing SQL queries
• Chart Generation: Visualize data with interactive charts using the datasette-agent-charts plugin
• Extensible Architecture: Add new functionality through plugins, similar to how Datasette operates
• Integration with Popular LLMs: Currently uses Gemini 3.1 Flash-Lite but designed to support others

The project’s demo video showcases these capabilities with real-world datasets like the Global Power Plants database from WRI and a backup of Willison’s own blog.

What This Means for Data Professionals in Africa:

Datasette Agent has particular promise for markets where data literacy may vary. The natural language interface lowers the barrier to entry, allowing more people to explore critical datasets related to development, health, education, and other key sectors.