Give your video a clear voice.

Start with a script you can say aloud.

For a simple creator-workflow video, try: “Give your next story room. Start with one clear idea, then shape it into a video worth sharing.” Read it at a comfortable pace and leave room for the images. Use the script timer as an estimate, then use the actual recording to set your scene durations.

Add a recording you own.

Open a video in Remixure, expand Audio and music, and confirm that you have permission to use the recording. Upload an MP3 or PCM WAV under 6 MB and ten minutes. Use Primary audio volume to balance the voice and Skip source intro (seconds) to skip a lead-in. Add a separate background music track when the scene needs it. Start quietly and make sure every word remains clear.

Audio begins at its chosen video entrance. A short recording leaves silence at the end; a long recording ends with the last scene. Play the preview from the beginning, and listen across scene changes.

Reuse a short video soundtrack.

In Audio and music, use Reuse a video soundtrack to read a local H.264/AAC MP4 up to 30 seconds and 6 MB. Choose Primary audio or Background music, and keep the decoded level or select a sample-peak ceiling. Prepare the WAV, compare both recordings, then confirm your rights and replacement review before choosing Use WAV. Only that separate WAV is uploaded; your MP4 stays on your device.

Changing the output requires another preparation. If the video changes, prepare again for its current draft. An unconfirmed upload retains the same WAV and request key: Retry this WAV checks that request. Check the private library before discarding. Save after use; Undo restores the earlier track and captions. Primary replacement clears captions and their appearance, while music replacement keeps them. Review the finished mix before approval.

If your recording is in an MP4, use the free video audio to WAV tool to read a permitted local H.264/AAC clip up to 30 seconds and 6 MB. Compare the original and separate WAV, optionally adjust its sample peak, then download and upload the reviewed WAV as primary audio. The original stays unchanged; this does not remove noise, certify loudness or update your project automatically.

Record a take in your browser.

In Audio and music, open Record your voice. Confirm consent from everyone being recorded and permission to use the take, then choose Start recording and allow microphone access in your browser. Record up to 60 seconds and choose Stop recording. Listen to the local playback before choosing Use as primary audio. Capture stops when you cancel, leave, hide the page or reach the limit; a take interrupted by the microphone needs an extra check of its ending.

The take stays local until you explicitly use it. Use uploads a private mono PCM WAV and replaces the primary track in your current draft. Save the video to retain the edit; Undo restores the previous audio and captions. A short take leaves silence after it ends. If upload confirmation is lost, Retry this take keeps the same recording and request key. Check the private library before starting another upload. Discard take releases local playback. Reloading or leaving loses an unused local take; audio already uploaded remains in the private library.

Recording needs a secure page and a browser with Web Audio recording support. If permission is denied or no microphone is available, allow access in browser settings or upload an MP3 or PCM WAV. Browser capture does not clone a voice, remove noise, guarantee studio quality or automatically create captions.

Or create a short synthetic narration.

Save the video, then open Create AI narration. Choose English or Spanish, then a voice. English offers Standard and 40 named presets; Spanish offers ten. Standard, Luna, Apollo, Aquila and Sirio have recorded previews of our original sentences. Playing a preview uses no trial attempt and does not read your script. Other presets clearly show that no recording is available; create a candidate and review its delivery before using it.

Enter up to 600 characters and confirm your script rights. The early-access trial allows five attempts per workspace per day across voices. A failed or cancelled attempt counts toward that limit. Availability also depends on shared service capacity; all named presets share a smaller daily service allowance across both languages. Changing the language or voice keeps your words and clears the script-rights checkbox, so check the new choice before creating it.

Listen before choosing Approve and use. Check names, numbers, abbreviations, pronunciation, pacing, and whether the voice suits the claim you are making. Approval adds the soundtrack to your draft; save again to keep that version. Write your script in the language you choose. This feature does not clone a person’s voice or translate your script. Automatic captions support English and Spanish. Save the approved narration, then create captions using its language. Review and correct the returned words and timing before applying them. You can also import an SRT or VTT track.

Help the voice say a name or abbreviation.

Use the free pronunciation spelling planner to compare wording before creating audio. Copy its draft JSON, then open Import narration wording in Create AI narration. Review and explicitly use it; this changes only that form and clears previous review and rights. Undo wording import restores the prior form while it remains unchanged.

In Create AI narration, use Add pronunciation when a written name needs a clearer spoken spelling. For example, keep “Remixure makes AI scripts easier to review.” as your original script and add Remixure → Remixure and AI → A I. The Spoken script preview shows exactly what goes to the voice: “Remixure makes A I scripts easier to review.” Your original scene text stays unchanged.

Each narration can retain up to twelve entries. Terms match whole words or phrases regardless of letter case, and the longest matching phrase wins. Replacements happen once; a replacement is not processed as another term. Check the match count, including entries with zero matches. Both the original and spoken scripts must fit 600 characters. Select Review spoken script, then confirm your rights. Editing the script, voice, spelling or saved video version requires another review.

Listen to the candidate before approving it. Spoken spellings guide the model and do not guarantee a particular pronunciation or tone. The saved candidate retains the original script, spoken script and entries for later inspection. Remove an entry or choose Clear spellings to return to the original wording; saved candidates remain unchanged. Use Refresh narrations to check service availability without discarding the current script or entries. These entries belong to that narration request and do not automatically correct later captions. Review names and abbreviations in the caption track too.

Check the result in context.

Watch the preview with sound and muted. Scene supporting lines follow their scenes. You can also import or create an editable timed caption track from the primary recording. Changing its start offset clears old captions so they cannot silently follow the wrong words; Undo restores them. You can download SRT or VTT and adjust a separate track with the caption editor. For the visual setting, create and review a scene image.

Save your final revision, then choose Export MP4. Choose 720p or 1080p when available: ordinary exports support up to two minutes at 720p or one minute at 1080p, both at 30 frames per second. Some footage and transition workflows have shorter limits shown in Check export. You can also download subtitles and the generated narration separately. A downloaded file is not a published social post.

Keep the recording honest.

Use only audio and words you have rights to use. Do not present a synthetic voice as a recording of a real person. Check the disclosure rules of the destination where you intend to share the video, and review product claims before posting.