frontend-plain

Plain HTML frontend — voice sample

Single-file vanilla HTML/CSS/JS — no build, no node_modules, no TypeScript. The actual file is at examples/voice/static/plain/index.html — that's both the source and what the HTTP-ingress singleton serves under /static/plain/.

To poke at it:

1. Start a voice cluster (one or more terminals — see top-level examples/voice/README.md).
2. Open http://localhost:8081/static/plain/.
3. Click "Enable mic" (one-time browser permission prompt + audio context unlock for inbound playback).
4. Log in with one of the test users from the selector page.
5. Try the three modes:
   • 1:1 — hold the PTT button next to a user's name; release to end the press.
   • Group — hold a group card's PTT; everyone in that group hears it (no echo of your own voice).
   • Room — enter a room, toggle "Talk" for open-mic. Multiple speakers stream concurrently; the binary frame envelope's senderUsername prefix lets the playback module dispatch each chunk to the right per-sender MediaSource.

The single-file layout is intentional: it makes the audio plumbing (capture + WebSocket I/O + per-sender playback) inspectable in one place without bouncing through framework abstractions. The other five frontends factor the same logic into framework-shaped pieces for comparison.