Bundled voice manifest loader.
Source: web/backend/services/voices.py
Reads a manifest.json file from web/backend/voices/ that defines bundled voice samples. Each voice has a reference audio file and metadata.
BundledVoice| Field | Type | Description |
|---|---|---|
id |
str |
Unique voice identifier |
name |
str |
Display name |
language |
str |
Language code |
description |
str |
Voice description |
file |
str |
Reference audio filename |
ref_text |
str |
Reference transcript |
| Property | Type | Description |
|---|---|---|
abs_path |
Path |
Absolute path to audio file |
exists |
bool |
Whether the audio file exists on disk |
| Function | Description |
|---|---|
load_voices() -> List[BundledVoice] |
Load all voices from manifest.json |
get_voice(voice_id) -> BundledVoice |
Look up a voice by ID |
Voices Service
├── manifest.json (voice definitions)
└──▶ consumed by Voices Router
└──▶ consumed by Pipeline Tasks (voice resolution)