Datasette Agent 0.5a0 adds side-panel chat for live app editing
The open-source Datasette Agent gains a 'panel' feature that lets an LLM assistant chat alongside Datasette pages and modify database-backed apps in place.
1 source · cross-referenced
- Datasette Agent 0.5a0 introduces a side-panel chat interface that runs alongside Datasette pages such as tables or queries.
- Users can now prompt an agent to create a database-backed app and iteratively edit it from the same chat panel.
- The feature integrates with datasette-apps to enable real-time app generation and modification.
- A live demo is available at agent.datasette.io with GitHub authentication.
The 0.5a0 release of Datasette Agent adds a "panel" feature that embeds a persistent chat alongside Datasette pages such as tables or query results. The chat remains open as users navigate, enabling continuous interaction with an LLM agent.
When paired with the datasette-apps plugin, the panel can display the app the agent has generated, with "View app" and "Edit" controls. Users can issue follow-up prompts like "give it a more hacker aesthetic design," and the agent will execute tool calls (app_create, app_render, app_view, app_str_replace) to update the app in place.
The release highlights a split-screen workflow: a narrow left column hosts the conversation, while the right panel shows the live app. For example, an assistant can build an "LLM Price Explorer" that displays 185 models from 12 providers with cost metrics, and a user can request style changes that the agent applies without leaving the interface.
To try the feature, users can sign in with GitHub at agent.datasette.io. The project is open source under the Datasette umbrella.
- Nov 11, 2026 · Latent Space — swyx
OpenAI demonstrates voice steering, cloud handoffs, and personal agents for Codex at AI Engineer Code Summit
Trust79 - Nov 11, 2026 · The Verge — AI
OpenAI to launch verified AI Business Agents in ChatGPT for US e-commerce and service businesses
Trust79 - Nov 10, 2026 · Latent Space — swyx
NaiveAI releases Naive-N0.5-Flash coding model with ultrafast token throughput
Trust74