# Teamspeak 6 Web Server Status
A simple webview of your Teamspeak 6 Server. This is a Python re-write of [ts6-web-explorer](https://github.com/Guraz/ts6-web-explorer) along with a visual overhaul 

<img width="852" height="906" alt="image" src="https://github.com/user-attachments/assets/00cf9ea7-cfad-4633-8d3b-7a66039fb87b" />

Visitors who know the server password can also **join voice straight from the browser**, no TeamSpeak client needed.

# Running
Requires Python 3.13+ (uv) and, for browser voice, Node.js 22.12+.
```
uv sync
uv run run.py
```
That one command starts the dashboard **and** the voice gateway (gateway npm dependencies are installed automatically on first run). Open `http://<host>:5000`.

- Copy `.env.template` to `.env` and fill in the secrets (`TS_HOST`, `TS_API_KEY`, `TS_SERVER_PASSWORD`)
- Everything else is configured in `config.toml`, which is shared by the dashboard and the voice gateway
- Images in `config.toml` are referenced based on the root of this repo
- Spacer channels are hidden
- Channels with the text `AFK` are shown, the number of users in this channel is used as the AFK users indicator
- `uv run app.py` still runs the dashboard alone (no voice)

# Browser voice
```
browser ──HTTP/WS :5000──▶ gateway (Node) ──proxy──▶ dashboard (Flask, 127.0.0.1:5001)
                              │
                              └──UDP voice──▶ TeamSpeak server :9987
```
- The gateway owns the public port: it serves `/ws/voice` and forwards everything else to Flask, so there is only one port to expose / reverse-proxy.
- Each browser gets its own TeamSpeak client inside the gateway (via [`@honeybbq/teamspeak`](https://github.com/HoneyBBQ/teamspeak-js), pure JS). Mic audio is sent as 20 ms PCM frames over the WebSocket and encoded to Opus on the gateway; incoming Opus is decoded in the browser with WebCodecs (or on the gateway for browsers without it). No WebRTC.
- **Access control:** voice is only offered when `TS_SERVER_PASSWORD` is set. Visitors must type the server password; the gateway checks it (constant time) *before* opening any TeamSpeak connection, and the TeamSpeak server checks it again. Repeated wrong passwords lock the IP out (`[voice] max_failed_logins`, `lockout_minutes`). Cross-origin WebSocket connections are rejected.
- Limits for small machines: `[voice] max_sessions` and `max_sessions_per_ip`.
- While connected, the channel list is rendered live from the voice session (talking indicators, join buttons per channel, password-protected channels prompt for their password).
- Controls: mute, deafen, voice activity (with sensitivity) or push-to-talk (hold the button or `V`), output volume.

## Deployment notes
- **Browsers only allow microphone access on HTTPS or `localhost`.** Put the app behind a TLS reverse proxy (Caddy, nginx, ...) and forward WebSocket upgrades for `/ws/voice`. If your proxy rewrites the `Host` header, add your public origin to `[voice] allowed_origins`; set `trust_proxy = true` so per-IP limits use `X-Forwarded-For`.
- **arm64 / SBCs:** everything is pure JS/WASM (Opus via `opusscript`), so `npm` needs no compiler. If you want lower CPU use, `cd gateway && npm install @discordjs/opus` (native addon) and the gateway will pick it up automatically.
- The gateway needs outbound UDP to the TeamSpeak voice port (`TS_VOICE_PORT`, default 9987).
- Gateway tests: `cd gateway && npm test`.