22 components in 3 categories
AI agent UI components
Agents do work over time, so their interface has to show progress, ask for approval and make it easy to step in. These components cover agent cards (an agent at work, its context window, a team to hire), run steps (a run timeline, a retrieval pipeline, plan then do, a progress ring with an ETA) and full agent workspaces.
The ten workspaces are among the largest pieces in the library: an approval inbox, an agent fleet board, a tool call inspector, a computer-use session, a run trace waterfall, handoff and escalation, scheduled agents and an agent builder canvas. Each is a complete screen with a phone layout, and its prompt rebuilds it inside your own product.
Each one is three small files (HTML, CSS and vanilla JS, no framework) and a prompt that rebuilds it in Claude, Cursor, v0 or Lovable. All of them are free with an account.