Sound that follows the edit
The agent marks every cut before any sound is made. Effects are described shot by shot and the music is planned around the edit, so the sound changes when the picture does.
Add sound to your video with AI. It watches the film, then scores it.
DFIRST Add sound to video is an AI tool that gives a film a new soundtrack. Upload an MP4, MOV or WebM of up to 60 seconds and choose sound effects, music, a voice-over or all three. DFIRST watches the film first, and you get the same video back with its sound replaced, each layer on its own track.
- Effects, music and voice-over, or just one
- Follows the cuts in your film
- Every layer adjustable afterwards
Sound that follows every cut.
Effects, music and a voice made for this film. Play it with the sound on.
How to add sound to a video with DFIRST.
Three steps. The first one starts on this page.
- 01
Add your film
Drop in a video of 1 to 60 seconds and choose what it needs: sound effects, music, a voice-over or any mix. Describe the sound or a music style to steer it, or leave both empty and let the agent decide from the picture.
- 02
DFIRST watches and scores it
An agent watches the film and writes a shot list with the cut points. Effects are made shot by shot, music is composed to fit, and the voice-over is written for the film or read from your text. Then everything is mixed under the original picture.
- 03
Fine-tune the mix
You get one video with the new soundtrack. Turn a track off, change its level, ask for a new version of one layer or type what to change, and only that part is made again.
Made for brands. Not a blank prompt box.
Every layer is its own track
Effects, music and voice come back as separate tracks. Lower the music, switch the effects off or ask for a new voice line without touching the rest. Each change makes a new version, and the earlier ones stay.
Speech on camera stays
When someone talks on camera, Auto picks Sonilo, which keeps their voice and replaces the rest of the sound. The picture itself is never changed: only the soundtrack is new.
Who uses it.
Brand and social teams
Finished sound for product films, reels and stories, without a sound designer on every project.
Agencies
A full soundtrack on a film from Storyboard to video or on an animatic, so the client hears the idea as well as sees it.
Video editors
A first sound pass on a cut: effects per shot and a music bed to refine or replace.
E-commerce teams
Sound and a short spoken line for silent product clips and demos.
At a glance.
The facts about Add sound to video, as it works in DFIRST today.
- Input
- One video: MP4, MOV or WebM, 1 to 60 seconds, up to 150 MB
- Layers
- Sound effects, music and voice-over, at least one
- Effects model
- Auto, Sonilo or Mirelo
- Your direction
- All optional: the sound (up to 2,000 characters), a music style (up to 1,000) and voice-over text (up to 1,500)
- Voice-over
- 8 voices; Polish, English, German, Spanish, French or Italian
- Models
- Gemini 3.8 Flash watches the film; Sonilo or Mirelo for effects; ElevenLabs Music v2.5 or Sonilo for music; ElevenLabs v4 for the voice
- You get
- The same video with a new soundtrack, every layer adjustable
Questions people ask.
About Add sound to video and how it fits the rest of DFIRST.
How do I add sound to a video with AI?
Upload your film in the box at the top of this page, describe the sound or leave it to the agent, and continue in DFIRST. Choose effects, music or a voice-over, and the video comes back with the new soundtrack mixed under the picture.
Can I add music and a voice-over to a video?
Yes. Switch on music, voice-over or both, next to the sound effects or instead of them. Describe a music style or let it follow the picture, and type the voice-over words or let the agent write a short line that fits the film.
Which sound effects model should I choose?
Leave it on Auto. The agent watches the video first: with several shots or someone speaking, it uses Sonilo, which follows the cuts and keeps the voice. For one continuous shot it uses Mirelo, which follows what happens on screen, such as steps, impacts and wind.
What happens to the original sound?
It is replaced by the new soundtrack, and the picture stays exactly as it was. If someone speaks on camera, Sonilo keeps their voice and drops the rest of the original sound. With Mirelo chosen by hand, the speech is not kept.
How long can the video be?
From 1 to 60 seconds, up to 150 MB, as an MP4, MOV or WebM file. For a longer film, add sound to each part of a minute or less and join them in your editor.
Which languages can the voice-over speak?
Polish, English, German, Spanish, French and Italian, in one of eight voices. Leave the language on "Match the video" and the line follows the text on screen, or English when there is none. For a voice-over on its own, use the AI voiceover generator.
Next steps. Tools that pick up where this one ends.
Start it here. Finish it in DFIRST.
Your add sound to video draft waits on this device until your access starts. DFIRST opens to people on the waitlist first.
Updated