Skip to content
AI Primer
release

Datasette Agent releases 0.1a3 with SQL chat, charts, and Fly sandbox plugins

Simon Willison shipped the first Datasette Agent release and companion chart and Fly sandbox plugins for conversational SQLite workflows. The stack combines live SQL inspection, chart rendering, and optional command execution inside an extensible local data assistant.

3 min read
Datasette Agent releases 0.1a3 with SQL chat, charts, and Fly sandbox plugins
Datasette Agent releases 0.1a3 with SQL chat, charts, and Fly sandbox plugins

TL;DR

  • simonw's launch thread introduced Datasette Agent as a first alpha for conversational Q and A over SQLite databases inside Datasette, with an interactive demo.
  • According to the main announcement post, the live demo runs on Gemini 3.1 Flash-Lite and can answer questions by writing SQL against example databases, then returning both the result and a Markdown export of the conversation.
  • the charts plugin release and the previous charts update show that chart generation shipped alongside the agent, with interactive tooltips, sequential and categorical color handling, and explicit execute-sql permission checks.
  • the sprites plugin release adds a separate plugin for running commands in a Fly Sprites sandbox, which turns the assistant into something broader than a SQL chat layer.
  • the 0.1a3 release note tightened the inspection loop with “View SQL query” buttons, better truncated-response handling, and cleaner reasoning output.

You can browse the official announcement, read Simon Willison's deeper write-up, and try the live demo. The release also arrived with a charts plugin and a Fly Sprites sandbox plugin, which makes the first alpha feel more like a small agent platform than a single feature.

SQL chat

Datasette Agent is a chat interface for data already stored in Datasette. In the main announcement post, Simon Willison walks through a prompt asking when he most recently saw a pelican; the system answers by generating a SQL query over his blog-backed SQLite data and returning the result.

The first alpha already exposes a few useful inspection hooks. the 0.1a3 release note says the UI now shows “View SQL query” buttons for visible tables and collapsed SQL result tool calls, and keeps table output visible even when the model response itself gets truncated.

Charts plugin

The charting layer shipped as its own plugin rather than being folded into the core agent. the charts plugin release says version 0.1a2 added “View SQL query” buttons below rendered charts, which keeps the chart output tied to the query that produced it.

datasette-agent-charts 0.1a2

Release: datasette-agent-charts 0.1a2 "View SQL query" buttons below rendered charts. Tags: datasette, datasette-agent

The previous day's earlier charts update filled in more of the mechanics:

  • Bar and waffle charts without a color column are shaded by magnitude.
  • Text-valued color columns use the observable10 categorical scheme.
  • Charts display interactive tooltips.
  • The plugin checks execute-sql permission before running its column-name lookup query.
  • A waffleY description bug was fixed so the agent can describe that chart type correctly.

That split is the interesting design choice here. The agent writes SQL, while chart rendering and chart-specific permissions live in an extension layer.

Fly sandbox plugin

A separate plugin called sprites pushes the project past read-only data chat. the sprites plugin release describes datasette-agent-sprites 0.1a0 as a Datasette Agent plugin for running commands in a Fly Sprites sandbox.

datasette-agent-sprites 0.1a0

Release: datasette-agent-sprites 0.1a0 A Datasette Agent plugin for running commands in a Fly Sprites sandbox. Tags: sandboxing, datasette, fly, datasette-agent

That matters because the launch thread framed Datasette Agent as extensible from day one. simonw's launch thread called out plugins for extra tools and features, and the same-day sprites release shows what that means in practice: one plugin adds charts, another adds sandboxed command execution.

Further reading

Discussion across the web

Where this story is being discussed, in original context.

·
Other sources· 3 posts

Datasette Agent

We just announced the first release of Datasette Agent, a new extensible AI assistant for Datasette. I've been working on my LLM Python library for just over three years now, and Datasette Agent represents the moment that LLM and Datasette finally come together. I'm really excited about it! Datasette Agent provides a conversational interface for asking questions of the data you have stored in Datasette. Add the datasette-agent-charts plugin and it can generate charts of your data as well. The demo The announcement post (on the new Datasette project blog) includes this demo video: I recorded the video against the new agent.datasette.io live demo instance, which runs Datasette Agent against example databases including the classic global-power-plants by WRI, and a copy of the Datasette backup of my blog. The live demo runs on Gemini 3.1 Flash-Lite - it's cheap, fast and has no trouble writing SQLite queries. A question I asked in the demo was: when did Simon most recently see a pelican? Which ran this SQL query: SELECT title, commentary, created FROM blog_beat WHERE beat_type = 'sighting' AND (title LIKE '%pelican%' OR commentary LIKE '%pelican%') ORDER BY created DESC LIMIT 5 And replied: The most recent sighting of a pelican by Simon was recorded on May 20, 2026. The observation included a California Brown Pelican, along with a Common Loon, Canada Goose, Striped Shore Crab, and a California Sea Lion. Here's that sighting on my blog, and the Markdown export of the full conversa

datasette-agent 0.1a3

Release: datasette-agent 0.1a3 "View SQL query" buttons for both visible tables and collapsed SQL result tool calls. Don't display empty reasoning chunks Improved handling of truncated responses - table still displays to the user even if the SQL results were truncated when showing the agent. See Datasette Agent, an extensible AI assistant for Datasette. Tags: datasette, datasette-agent

datasette-agent-charts 0.1a1

Release: datasette-agent-charts 0.1a1 More color! Bar and waffle charts without a color column are shaded by magnitude with a sequential color scheme; color columns holding text values use the observable10 categorical scheme. #2 Now checks execute-sql permission before running the query to find the column names. Charts now display interactive tooltips. Fixed a bug where waffleY charts were not described to the agent. Tags: datasette, datasette-agent

Share on X