# VOX TTS Server

A standalone HTTP server that produces Half-Life / Black Mesa VOX announcer
speech by concatenating pre-recorded word clips. It is based on VOXGen by
wphillips (https://github.com/wphillips/VOXGen).

This service was extracted from the ttstoy Discord bot cog so it can run on its
own as a system service, independent of the bot.

## How it works

The input text is split into words, and each word is matched to a `.wav` clip in
the selected voice pack. Matching clips are concatenated into a single WAV
response. Words with no matching clip are skipped.

Two packs are bundled:

- `vox`: the classic Half-Life VOX announcer words (in `vox_words/`)
- `vox2`: the Black Mesa announcement system words (in `vox2_words/`)

## Requirements

- Python 3.9 or newer
- No third-party packages: the server uses only the Python standard library

## Running

```bash
./start.sh
```

The listening port defaults to `33003` and can be overridden with the `PORT`
environment variable.

You can also run it directly:

```bash
PORT=33003 python3 server.py
```

## API

### GET /say

Generate speech and return a WAV file.

```
GET /say?text=warning+reactor+core+meltdown&pack=vox
```

Query parameters:

- `text` (required): words to synthesize, separated by spaces
- `pack` (optional): `vox` (default) or `vox2`

Response: `audio/wav` (WAV bytes).

Example:

```bash
curl "http://127.0.0.1:33003/say?text=alert+intruder&pack=vox" -o vox.wav
```

### GET /packs

Returns a comma-separated list of installed packs.

### GET /health

Returns `ok`.

## Running as a system service

A systemd unit file is provided (`vox-tts-server.service`). See INSTALL.md for
setup steps.