Sirene

Appearance

API Server (Hono) ​

The Hono server runs on port 3000 and acts as the orchestrator. It handles validation, resolves voices/parameters, and forwards inference requests to the FastAPI service.

Generation ​

Generate Speech ​

http
POST /api/generate

Body:

json
{
  "model": "kokoro-v1",
  "input": "Hello, world!",
  "voice": "voice_id",
  "speed": 1.0,
  "language": "en",
  "stream": false
}

Response: audio/wav or audio/pcm (when stream: true)

Voices ​

List Voices ​

http
GET /api/voices

Create a Voice ​

http
POST /api/voices

Body: multipart/form-data with name, description, language, avatar.

Get Voice Details ​

http
GET /api/voices/:id

Update a Voice ​

http
PUT /api/voices/:id

Delete a Voice ​

http
DELETE /api/voices/:id

Add a Sample ​

http
POST /api/voices/:id/samples

Body: multipart/form-data with audio file and transcript.

Delete a Sample ​

http
DELETE /api/voices/:id/samples/:sid

Models ​

Get Model Catalog ​

http
GET /api/models/catalog

Returns the list of available models from embedded manifests.

Get Installed Models ​

http
GET /api/models/installed

Scans the filesystem to list installed models.

Pull a Model ​

http
GET /api/models/:id/pull

Response: Server-Sent Events stream with download progress:

data: {"progress": 45, "speed": "12.5 MB/s", "status": "downloading"}
data: {"progress": 100, "status": "complete"}

Delete a Model ​

http
DELETE /api/models/:id

Model Events ​

http
GET /api/models/events

SSE stream notifying filesystem changes (model installed/removed).

Transcription ​

Transcribe Audio ​

http
POST /api/transcribe

Body: multipart/form-data with audio file.

Response:

json
{
  "text": "transcribed text",
  "language": "en"
}

History ​

List Generations ​

http
GET /api/generations

Paginated list with filters: voice, model, language, date.

Get Generation Details ​

http
GET /api/generations/:id

Delete a Generation ​

http
DELETE /api/generations/:id

System ​

Health Check ​

http
GET /api/health

Version ​

http
GET /api/version