Generate studio-quality voiceovers directly in Mirage Studio, or upload your own audio. This guide covers prompting for expressive delivery, choosing the right voice, supported languages, voice cloning, and limits.

​
Prompting guide: Enhancing emotion

Use inline audio tags and prompt techniques to direct performance.
  • Emotional tags: Add tags like [excited], [calm], [sad], [whispers], [laughs] inline with your text to shape delivery.
  • Punctuation: Use ! for energy, ... for pauses, ? for rising intonation, , for brief pauses, : to cue anticipation, and ; for longer pauses.
  • Emphasis: WRITE words in ALL CAPS to increase emphasis and pitch.
See full guides:

​
Manually adjusting emotion

You can refine delivery without changing your script:
  • Insert or remove emotion tags around phrases to increase or reduce intensity.
  • Add pacing with ellipses (...) and commas to slow down or breathe between ideas.
  • Regenerate just the line/segment after small prompt tweaks for fast iteration.
Example: “We did it [excited]! … But [calm] there’s more work ahead.”

​
Voice selection: Male vs. Female

Choose a voice that matches the on-screen actor to maintain realism and lip-sync believability.
  • Match appearance: If the actor presents as female, choose a female voice; if male, choose a male voice.
  • Context matters: For narration over b‑roll, you can choose any voice; for talking-head shots, match gender and tone to the actor.
  • Brand consistency: Keep the same voice across versions unless testing variants.

​
Languages

  • Text-to-speech languages: Arabic, Chinese (Simplified), 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.
  • Uploading audio: All languages are supported if you upload your own recorded audio; Mirage will use the language present in the file.
If your language isn’t supported in TTS yet, record or upload your own track to proceed today.

​
Voice clone

Create a voice that sounds like you or your brand’s talent.

​
Create a voice clone

  1. Open the voice actors section, then click on Create voice clone.
  2. Record your voice for around 30 seconds (with energized voice and minimum background noise).
  3. Preview the result and save or discard.

​
Character & time limit

  • Character limit: Your prompt should be at least 70 characters and maximum 2000 characters long (which is equal to 60 seconds).
    • Note: Free users have a 15 second video generation limit which is equal to about 500 characters.
If you hit a limit, split your script into multiple segments and generate per segment. Then assemble in the timeline.