GLM-TTS

POST

/bigmodel/api/paas/v4/audio/speech

The GLM-TTS speech synthesis model, with the new-generation Zhipu speech large model at its core, breaks through the traditional speech synthesis framework. It intelligently predicts the emotion and intonation of the text through context, significantly improving the naturalness and expressiveness of the speech, making the synthesized speech possess real emotions and vitality. GLM-TTS adopts a two-stage generation in its architecture and introduces a reinforcement learning scheme based on GRPO during training, achieving open-source SOTA performance in the publicly evaluated "character error rate" and "emotional expression".

Price: 0.03 PTC/1000 characters

Request

Query Params

Header Params

Body Params application/json

Example

{
      "model": "glm-tts",
      "input": "Hello, welcome to the Zhipu Open Platform",
      "voice": "tongtong",
      "response_format": "wav",
      "speed": 1.0,
      "volume": 1.0
}

Request Code Samples

Shell

JavaScript

Java

Swift

PHP

Python

HTTP

Objective-C

Ruby

OCaml

Dart

curl --location --request POST 'https://api.302.ai/bigmodel/api/paas/v4/audio/speech?output_format=url' \
--header 'Authorization: Bearer ' \
--header 'Content-Type: application/json' \
--data-raw '{
      "model": "glm-tts",
      "input": "Hello, welcome to the Zhipu Open Platform",
      "voice": "tongtong",
      "response_format": "wav",
      "speed": 1.0,
      "volume": 1.0
}'

Responses

🟢200success

application/json

Body

Example

{
    "url": "https://file.302ai.cn/gpt/imgs/20251216/8611be2603f28f401aeeb1bd2b115fae.wav"
}

Modified at 2025-12-16 03:34:49

wavespeed-ai/infinitetalk-fast/video-to-video

GLM-TTS-Clone