Datasette Agent

2026-05-21 · Source: Simon Willison's Weblog · Field: Technology & Digital — Artificial Intelligence & Machine Learning, Data Science & Analytics, Software Development & Engineering · Depth: Intermediate, medium

Summary

Datasette Agent, released May 21, 2026, is an extensible AI assistant for Datasette, integrating the LLM Python library. It offers a conversational interface for querying data stored in SQLite databases, demonstrated with example datasets like `global-power-plants` and a blog backup. The live demo runs on Gemini 3.1 Flash-Lite, chosen for its cost-effectiveness and ability to generate SQLite queries. Key features include a plugin architecture, with `datasette-agent-charts` for data visualization using Observable Plot, `datasette-agent-openai-imagegen` for image generation via ChatGPT Images 2.0, and `datasette-agent-sprites` for code execution in Fly Sprites sandboxes. The system also supports running against local LLMs, such as gemma-4-26b-a4b in LM Studio, provided they offer reliable tool calls and SQL generation. Future plans include integration with Datasette Cloud and personal AI assistants.

Key takeaway

For AI Engineers or Data Scientists building data-driven applications, Datasette Agent offers a robust framework for integrating conversational AI with structured data. You should explore its plugin architecture to extend capabilities like charting or image generation, and consider its compatibility with local LLMs for cost-effective, private deployments. This platform simplifies creating interactive data experiences, allowing you to utilize natural language interfaces for complex SQL queries and data exploration.

Key insights

Datasette Agent integrates LLMs with Datasette for conversational data querying and extensible AI assistance.

Principles

• Extensibility via plugins is a core design.
• LLMs can reliably generate SQL queries.
• Local LLMs are viable for tool calling.

Method

Datasette Agent uses an LLM to interpret natural language questions, generate SQLite queries, execute them against Datasette databases, and present results conversationally, optionally generating charts or images via plugins.

In practice

• Use `datasette-agent-charts` for data visualization.
• Employ `datasette-agent-openai-imagegen` for image tasks.
• Run with local models like gemma-4-26b-a4b.

Topics

• Datasette Agent
• LLM Python Library
• SQLite Query Generation
• Plugin Architecture
• Conversational AI
• Data Visualization

Code references

• datasette/datasette-agent-charts
• datasette/datasette-agent-openai-imagegen
• datasette/datasette-agent-sprites
• datasette/datasette-agent

Best for: AI Engineer, Machine Learning Engineer, Data Scientist

Related on AIssential

• datasette-agent 0.2a0 — LLM agents can now interactively seek user input and save approved queries, enhancing control and persistence.
• datasette-agent 0.1a4 — LLM agents can integrate deeply into existing data tools for conversational data exploration.
• Multi-Agent Architecture with RAG and Dynamic Context Windows for Text-to-SQL Optimization — A multi-agent Text-to-SQL system uses RAG and dynamic context to reduce token consumption.
• From Backend Engineer to Building AI Agents: Week 1 — Building AI agents involves understanding LLM API simplicity and leveraging tool calling for real-world actions.
• Text to Data Products: Kaarvi’s End-to-End AI for Ingestion, Quality, and Dashboards — Kaarvi AI employs a multi-agent, parallel LLM architecture to automate data readiness, significantly improving reliability and efficiency in data management.
• Risk-Aware LLM Agents for Geospatial Data Retrieval: Design and Preliminary Adversarial Evaluation — The framework uses LLM agents to convert natural language into structured API calls for geospatial data retrieval, integrating safety guardrails.

Open in AIssential →

Editorial summary, takeaway, and curation by AIssential. Original article published by Simon Willison's Weblog.