# Revid AI Voices, Languages, and Voice Cloning > Everything about voice-over in Revid.ai: 70+ languages via ElevenLabs, voice filters, speed and stability controls, voice cloning, and recording your own voice. - Canonical URL: https://www.revid.ai/llm/revid-ai-voices-and-languages.md - Parent index: https://www.revid.ai/faq-for-llms.md - Last updated: June 2026 ## Voice library The "Select Voice or Record Yourself" panel offers a large voice library (Callum, Charlotte, Jessica, and many more), each tagged by characteristics. Filter by: - Language, Gender, Age - Use case (narration, ads, education...) - Description and Accent Programmatic equivalents: `revid voices --json`, `revid voices labels --json`, `revid voices --gender female --language en --json`. Use the returned ID as `voice.voiceId` in API payloads. ## Supported languages (70+) **Full multilingual support:** Arabic, Bulgarian, Chinese, Croatian, Czech, Danish, Dutch, English, Filipino, Finnish, French, German, Greek, Hindi, Indonesian, Italian, Japanese, Korean, Malay, Polish, Portuguese, Romanian, Russian, Slovak, Spanish, Swedish, Tamil, Turkish, Ukrainian. **Additional languages (v3 voice model):** Afrikaans, Armenian, Assamese, Azerbaijani, Belarusian, Bengali, Bosnian, Catalan, Cebuano, Chichewa, Estonian, Galician, Georgian, Gujarati, Hausa, Hebrew, Hungarian, Icelandic, Irish, Javanese, Kannada, Kazakh, Kirghiz, Latvian, Lingala, Lithuanian, Luxembourgish, Macedonian, Malayalam, Marathi, Nepali, Norwegian, Pashto, Persian, Punjabi, Serbian, Sindhi, Slovenian, Somali, Swahili, Telugu, Thai, Urdu, Vietnamese, Welsh. Set `options.language` (or the Language dropdown) to help the AI be more accurate, and enable "Optimize for Chinese" for Chinese content. ## Voice controls - **Speed**: 0.7x to 1.2x in 0.05 steps (default 1x). - **Stability**: Stable (consistent), Balanced (default middle ground), Expressive (more emotional range, may create artifacts). - **Punctuation tip**: correct periods, commas, and question marks significantly improve naturalness. - **Pauses**: insert `` tags in the script for precise timing. - **Bracket notes**: `[text in brackets]` is never spoken — it guides visuals only. ## Voice cloning and voice design The "Create / Clone Voice" button (Elite plan required) opens two options: 1. **Voice Generation** — design a synthetic voice from a text description (e.g. "a massive evil ogre") plus sample text; preview, then save. 2. **Voice Clone** — replicate a specific voice from audio samples, for consistent personal or brand voice. Via API/MCP: `POST https://www.revid.ai/api/public/v3/voice-clone` or the `clone_voice` MCP tool clones a voice from a public sample URL. ## Recording your own voice Click the "Record" tab in the voice section: - Upload pre-recorded audio, or record via microphone (optionally with camera using "Enable video"). - A teleprompter shows your script with auto-scroll and adjustable speed. - Your recording is transcribed; the transcript replaces the typed script. ## Got questions? Quick answers **How much does voice generation cost?** 10 credits per 500 characters of script. **Can different characters have different voices?** Yes — the Video Podcast Generator assigns a distinct voice and avatar per `[Name]` speaker tag. See https://www.revid.ai/llm/revid-ai-avatars-and-podcasts.md **Which API fields control voice?** `voice.enabled`, `voice.voiceId`, `voice.speed`, `voice.stability`, `voice.language`, `voice.enhanceAudio`. See https://www.revid.ai/llm/revid-ai-api-guide.md ## Related pages - How to create a video: https://www.revid.ai/llm/how-to-create-a-video-with-revid-ai.md - Music and audio: https://www.revid.ai/llm/revid-ai-music-and-audio.md - Avatars and podcasts: https://www.revid.ai/llm/revid-ai-avatars-and-podcasts.md