Updated at October 31, 2023 06:09 AM
Speech synthesis
The speech synthesis function allows you to voice strings of text. To synthesize speech, use a POST or GET request to https://voice.mcs.mail.ru/tts.
GET
POST
To synthesize speech using a GET request, send the text in the text parameter.
Request example:
curl -L --request GET \--url 'https://voice.mcs.mail.ru/tts?text=<text to voice>' \--output <file name>--header 'Authorization: Bearer <access token>'
Request parameters:
Parameter | Description | Transmission method | Accepted values |
|---|---|---|---|
model_name | Name of the voice model | URL | katherine (or katherine-hifigan) — by default; maria (or maria-serious); pavel (or pavel-hifigan) |
encoder | Encoder type | URL | pcm — by default; mp3; opus |
tempo | Speech rate | URL | 0.75 to 1.75 |
text | Text for voiceover (GET request) | URL | Up to 2000 symbols in UTF-8 |
<text to voice> | Text for voiceover (POST request) | Option data | Up to 2000 symbols in UTF-8 |
<file name> | File name with extention | Option output | — |
<access token> | Access token for authorization. Read more about authorization in the section Getting an access token | Option header | — |
The text passed in the request must be encoded in UTF-8 and contain no more than 20,000 characters.
The response contains an audio file of the selected format. By default, a PCM file is created.
Code | State | Description |
|---|---|---|
4048 | 400 | Invalid token |
4049 | 400 | Inactive VK Cloud project |
4051 | 14001 | Invalid text format |