Prompt Nodes Agent Runs Live Output

Run saved prompts against local apps from a visible graph.

Each node stores a prompt and working directory. Click a node to edit it, run it through Agent Manager, and inspect the live terminal output in the side panel.

Current model

Start with one-node prompt execution. Add more nodes only when you actually want a visible dependency graph instead of a single opaque session.

Code Graph

Ready to run Active run Completed Failed Connections are optional. Drag cards to arrange them, then click a node to edit or run it.