Voice Generation Essentials
Generate natural-sounding speech from text with six AI voices, ten languages, and adjustable playback speed. Preview inline, download the audio file, or compose it directly onto a video.
Step 1: Choose a Voice & Language
CreativeAI offers six distinct voice presets, each with a different tone and character. Pair any voice with one of ten supported languages.
Available voices: alloy β Neutral, balanced echo β Warm, conversational fable β Expressive, storytelling onyx β Deep, authoritative nova β Friendly, energetic shimmer β Soft, calm Supported languages: en English zh Chinese (Mandarin) ja Japanese ko Korean es Spanish fr French de German pt Portuguese ru Russian ar Arabic
Tip: nova and echo work well for marketing narration; onyx is ideal for documentary-style voiceovers.
Step 2: Write or Paste Your Script
Enter up to 5,000 characters of text. The API request is straightforward:
{
"text": "Welcome to CreativeAI. Transform your scripts into professional voiceovers in seconds.",
"voice": "nova",
"language": "en",
"speed": 1.0
}The response includes the audio URL once generation completes:
{
"id": "gen_abc123",
"status": "COMPLETED",
"output_url": "https://cdn.creativeai.run/audio/abc123.mp3",
"credits": 8,
"created_at": "2026-03-27T12:00:00Z"
}Step 3: Adjust Playback Speed
Control pacing with the speed parameter. Values range from 0.5Γ (slow, deliberate) to 2.0Γ (fast, energetic). The default is 1.0Γ.
0.5Γ
Audiobooks, meditation
1.0Γ
Default, natural pace
2.0Γ
Ads, short-form video
Step 4: Preview, Download & Compose
Preview β The Audio Studio includes an inline player. Hit play to hear the result before downloading.
Download β Save the generated MP3 for use in any editing tool or DAW.
Compose with Video β Use the compose endpoint to merge your audio onto a video track, replacing or mixing the original audio.
Compose request β merge audio onto an existing video:
{
"video_url": "https://cdn.creativeai.run/videos/my-video.mp4",
"audio_url": "https://cdn.creativeai.run/audio/abc123.mp3",
"replace_audio": true
}Response (poll for status until COMPLETED):
{
"id": "comp_xyz789",
"status": "PROCESSING",
"output_url": null,
"credits": 5,
"created_at": "2026-03-27T12:01:00Z"
}