# Local Text to speech server for TypeAgent shell
This directory contains examples of local Text to speech (TTS) server that works with the [TypeAgent shell](../../ts/packages/shell/). Additional TTS service for TypeAgent Shell can be built following the [REST API](#rest-api) description below.
## Usage
- Go to one of the TTS directory
- Follow the direction in the directory to setup .venv and start the service
- In [TypeAgent shell](../../ts/packages/shell/), change the TTS provider to local and choose a voice.
## REST API
The local TTS provider in TypeAgent expects two REST API:
### `GET /voices`
Return the available voices a JSON array of strings (`string[]`) or array of name, value string pair (`[string, string][]`).
### `POST /synthesize`
Body is a JSON object with the following typescript schema
```typescript
{
text: string
voiceName?: string
}
```
`text` - the text to synthesize speech for
`voiceName` - an optional voice name value from the `GET /voices`. The services will pick a default voice if it is missing.
Return a blob containing the synthesized speech in WAV format.microsoft/TypeAgent
Publicmirrored fromhttps://github.com/microsoft/TypeAgentAvailable
python/tts/README.md
33lines · modepreview