Agent Activity
Watch one agent run go from thinking to a finished plan, built from the step components.
This run puts the step components together the way a host app would. The agent thinks with Reasoning, calls two Tools at once, proposes a Plan, hands the work to a Subagent and keeps follow-up prompts in a Queue. A Status badge shows where the whole run stands.
Running
- Document the retry delay
- Write the changelog entry
- Draft the release notes
What Happens
- Thinking: Reasoning streams its text while
running, then closes and reads "Thought for 6 s". - Tools: both calls run side by side.
delete_branchshows "Waiting for approval", and the run's badge reads "Needs input", until the host records the decision. The tool then keeps "Approved", not the request. - Plan: the run stops on the plan until you press Accept. Its steps then tick off as the subagent finishes them.
- Subagent: its own tool calls nest inside it, and its result shows once it's done.
- Queue: the follow-ups wait their turn. Drag one to reorder it, or focus it and press Space to pick it up, the arrow keys to move it and Space to drop it. Delete removes it.
Each component gets its state from the host as plain data, here a timer stepping through a script. In an app, the same data comes from your model's stream, for example through the AI SDK adapter.
Every component here runs in React and Web Components. The example is React.