AUDIO

🔊 Gemini TTS Model

Making Request:

Endpoint:

https://api.hyprlab.io/v1beta/models/[MODEL]:generateContent?key=[API-KEY]

Single-Speaker Example:

#!/bin/bash
set -e -E

HYPRLAB_API_KEY="$HYPRLAB_API_KEY"
MODEL_ID="gemini-2.5-flash-preview-tts"

curl -X POST \
-H "Content-Type: application/json" \
"https://api.hyprlab.io/v1beta/models/${MODEL_ID}:generateContent?key=${HYPRLAB_API_KEY}" \
-d '{
    "contents": [
      {
        "role": "user",
        "parts": [
          {
            "text": "Greet the user in a warm and welcoming voice: Good morning! Ready to start your day?"
          }
        ]
      }
    ],
    "generationConfig": {
      "responseModalities": ["audio"],
      "temperature": 1,
      "speech_config": {
        "voice_config": {
          "prebuilt_voice_config": {
            "voice_name": "Zephyr"
          }
        }
      }
    }
}' > output.json

echo "Response saved as output.json"

const axios = require("axios");
const fs = require("fs");

// API endpoint and headers
const modelId = "gemini-2.5-flash-preview-tts";
const url = `https://api.hyprlab.io/v1beta/models/${modelId}:generateContent?key=${process.env.HYPRLAB_API_KEY}`;
const headers = {
  "Content-Type": "application/json",
};

// Request payload
const data = {
  contents: [
    {
      role: "user",
      parts: [
        {
          text: "Greet the user in a warm and welcoming voice: Good morning! Ready to start your day?",
        },
      ],
    },
  ],
  generationConfig: {
    responseModalities: ["audio"],
    temperature: 1,
    speech_config: {
      voice_config: {
        prebuilt_voice_config: {
          voice_name: "Zephyr",
        },
      },
    },
  },
};

// Make the request
axios
  .post(url, data, { headers })
  .then((response) => {
    // Save the full JSON response to output.json
    fs.writeFileSync("output.json", JSON.stringify(response.data, null, 2));
    console.log("Response saved as output.json");
  })
  .catch((error) => {
    console.error("Error:", error.response ? error.response.data : error.message);
  });

import requests
import json
import os

# API endpoint and headers
model_id = "gemini-2.5-flash-preview-tts"
url = f"https://api.hyprlab.io/v1beta/models/{model_id}:generateContent?key={os.getenv('HYPRLAB_API_KEY')}"
headers = {
    "Content-Type": "application/json",
}

# Request payload
data = {
    "contents": [
        {
            "role": "user",
            "parts": [
                {
                    "text": "Greet the user in a warm and welcoming voice: Good morning! Ready to start your day?"
                }
            ]
        }
    ],
    "generationConfig": {
        "responseModalities": ["audio"],
        "temperature": 1,
        "speech_config": {
            "voice_config": {
                "prebuilt_voice_config": {
                    "voice_name": "Zephyr"
                }
            }
        }
    }
}

# Make the request
response = requests.post(url, json=data, headers=headers)

# Save the full JSON response to output.json
with open("output.json", "w") as f:
    json.dump(response.json(), f, indent=2)
print("Response saved as output.json")

Multiple-Speaker Example:

#!/bin/bash
set -e -E

HYPRLAB_API_KEY="$HYPRLAB_API_KEY"
MODEL_ID="gemini-2.5-flash-preview-tts"

curl -X POST \
-H "Content-Type: application/json" \
"https://api.hyprlab.io/v1beta/models/${MODEL_ID}:generateContent?key=${HYPRLAB_API_KEY}" \
-d '{
    "contents": [
      {
        "role": "user",
        "parts": [
          {
            "text": "Read aloud in a warm, welcoming tone\nSpeaker 1: Hello! We are excited to show you our native speech capabilities.\nSpeaker 2: Where you can direct a voice, create realistic dialog, and so much more. Edit these placeholders to get started."
          }
        ]
      }
    ],
    "generationConfig": {
      "responseModalities": ["audio"],
      "temperature": 1,
      "speech_config": {
        "multi_speaker_voice_config": {
          "speaker_voice_configs": [
            {
              "speaker": "Speaker 1",
              "voice_config": {
                "prebuilt_voice_config": {
                  "voice_name": "Zephyr"
                }
              }
            },
            {
              "speaker": "Speaker 2",
              "voice_config": {
                "prebuilt_voice_config": {
                  "voice_name": "Puck"
                }
              }
            }
          ]
        }
      }
    }
}' > output.json

echo "Response saved as output.json"

const axios = require("axios");
const fs = require("fs");

// API endpoint and headers
const modelId = "gemini-2.5-flash-preview-tts";
const url = `https://api.hyprlab.io/v1beta/models/${modelId}:generateContent?key=${process.env.HYPRLAB_API_KEY}`;
const headers = {
  "Content-Type": "application/json",
};

// Request payload
const data = {
  contents: [
    {
      role: "user",
      parts: [
        {
          text: "Read aloud in a warm, welcoming tone\nSpeaker 1: Hello! We are excited to show you our native speech capabilities.\nSpeaker 2: Where you can direct a voice, create realistic dialog, and so much more. Edit these placeholders to get started.",
        },
      ],
    },
  ],
  generationConfig: {
    responseModalities: ["audio"],
    temperature: 1,
    speech_config: {
      multi_speaker_voice_config: {
        speaker_voice_configs: [
          {
            speaker: "Speaker 1",
            voice_config: {
              prebuilt_voice_config: {
                voice_name: "Zephyr",
              },
            },
          },
          {
            speaker: "Speaker 2",
            voice_config: {
              prebuilt_voice_config: {
                voice_name: "Puck",
              },
            },
          },
        ],
      },
    },
  },
};

// Make the request
axios
  .post(url, data, { headers })
  .then((response) => {
    // Save the full JSON response to output.json
    fs.writeFileSync("output.json", JSON.stringify(response.data, null, 2));
    console.log("Response saved as output.json");
  })
  .catch((error) => {
    console.error("Error:", error.response ? error.response.data : error.message);
  });

import requests
import json
import os

# API endpoint and headers
model_id = "gemini-2.5-flash-preview-tts"
url = f"https://api.hyprlab.io/v1beta/models/{model_id}:generateContent?key={os.getenv('HYPRLAB_API_KEY')}"
headers = {
    "Content-Type": "application/json",
}

# Request payload
data = {
    "contents": [
        {
            "role": "user",
            "parts": [
                {
                    "text": "Read aloud in a warm, welcoming tone\nSpeaker 1: Hello! We are excited to show you our native speech capabilities.\nSpeaker 2: Where you can direct a voice, create realistic dialog, and so much more. Edit these placeholders to get started."
                }
            ]
        }
    ],
    "generationConfig": {
        "responseModalities": ["audio"],
        "temperature": 1,
        "speech_config": {
            "multi_speaker_voice_config": {
                "speaker_voice_configs": [
                    {
                        "speaker": "Speaker 1",
                        "voice_config": {
                            "prebuilt_voice_config": {
                                "voice_name": "Zephyr"
                            }
                        }
                    },
                    {
                        "speaker": "Speaker 2",
                        "voice_config": {
                            "prebuilt_voice_config": {
                                "voice_name": "Puck"
                            }
                        }
                    }
                ]
            }
        }
    }
}

# Make the request
response = requests.post(url, json=data, headers=headers)

# Save the full JSON response to output.json
with open("output.json", "w") as f:
    json.dump(response.json(), f, indent=2)
print("Response saved as output.json")

LEARN MORE: https://aistudio.google.com/generate-speech | https://ai.google.dev/gemini-api/docs/speech-generation

Pricing:

Model Name:

gemini-2.5-pro-preview-tts

Discount:

60% off

Input:

$0.4 / 1M Tokens

Output:

$8 / 1M Tokens

Context Length:

32,000

Moderation:

Unfiltered

Capabilities:

Single-Speaker Audio
Multi-Speaker Audio

Model Name:

gemini-2.5-flash-preview-tts

Discount:

60% off

Input:

$0.2 / 1M Tokens

Output:

$4 / 1M Tokens

Context Length:

32,000

Moderation:

Unfiltered

Capabilities:

Single-Speaker Audio
Multi-Speaker Audio

PreviousGoogle NextCHAT

Last updated 2 months ago