Verify each demo app works standalone, by voice, and driven by the assistant #248
Labels
No labels
meeting-notes
meeting-transcript
No milestone
No project
No assignees
1 participant
Notifications
Due date
No due date set.
Dependencies
No dependencies set.
Reference
lhumina_code/home#248
Loading…
Add table
Add a link
Reference in a new issue
No description provided.
Delete branch "%!s()"
Deleting a branch is permanent. Although the deleted branch may continue to exist for a short time before it actually gets removed, it CANNOT be undone in most cases. Continue?
Once a fresh tester provisions cleanly from the stable release, confirm the demo apps are actually usable end to end, in three layers. (1) Each app works on its own: planner, slides, whiteboard, the voice widget, and the Kimi assistant each open, render, and perform their core action. (2) Voice works across the apps where the widget is present: speak a request and hear the answer. (3) The assistant, driven by voice or text, operates the apps through its tools, for example create a board or a plan and have it appear in Planner or Whiteboard. The deliverable is a confirmed matrix of app, app-with-voice, and app-with-voice-and-assistant before wider onboarding. Parent: home#239.
For the by-voice and assistant-driven part on the planner app: the planner already has an assistant widget that talks to the orchestrator service, so to get a working live voice agent there we need to deploy the latest hero_voice from its main branch (it gained voice activity detection) and the latest hero_orchestrator from main, both onto the tester. The widget then picks up the orchestrator automatically. The shared voice engine on the admin machine is unchanged and already running. The assistant in the widget runs on the Kimi model with a limited tool set, which is fine for this. We will verify the speak to planner and respond flow on the tester after deploying.
Signed-by: mik-tf mik-tf@noreply.invalid