You can create a standalone voiceover in your browser. It does not require the API or a video project. Use AI Voice Generator for a preset voice, or AI Voice Cloner for a voice reference you have permission to use.
Generate speech from a script
Open AI Voice Generator and sign in.
Enter your script directly in the text area under Enter your script and select a voice. Clicking it lets you type or paste; a separate writing window is not required.
Choose a voice. Use Explore All to browse the available choices.
Set the language to match the script or use Auto. Adjust Speed if needed.
Review the credit estimate and select Generate.
Play the result, check the words and pronunciation, and download it from the completed result or the Audio section of My Library.
Clone a voice
Open AI Voice Cloner.
Upload a clear reference recording, or use Record audio after allowing microphone access. The current recorder recommends 10–30 seconds for best results; this recommendation is not a universal upload limit.
Enter the exact words you want the new voice to say.
Choose the language and speed, review the estimate, and select Generate Speech.
Voice cloning generates new speech from the script and voice reference. It is not a word-for-word audio replacement that guarantees the original timing, rhythm, emotion, pauses, or mixed-language pronunciation. Listen to a short test before creating a long passage. Use voices and source material only with the necessary rights and consent.
Languages and pronunciation
The current selector includes Auto, Chinese, English, Japanese, Korean, German, French, Russian, Portuguese, Spanish, and Italian. Auto is not a guarantee of support for every language. For names or mixed-language text, listen to the result and try shorter passages or clearer spellings. Changing the voice language does not translate the script automatically.
Use the audio in a video
Download the voiceover and add it to Talking Photo for a still portrait, Lip Sync for an existing video, or a video model that supports uploaded audio. See using your own audio in a generated video.
Troubleshooting
Cannot type: click or tap inside the script area itself. Save your text elsewhere before refreshing.
Generate unavailable: check the script, selected voice or reference recording, completed upload, and displayed balance.
Words are skipped, cut off, or changed: compare the submitted script with the generated audio. Note the exact timestamp and selected language. Stop repeated retries if the same problem continues.
For a supported upload or broken output, send support the tool, project link if available, language, visible error, and a short description of the problem. Never send passwords or access tokens.