Add sound to a silent video
See the featuretendre.AI can add a synchronized ambience track to a video that has no sound, using the local MMAudio model. It runs entirely on your own GPU, no cloud step involved.
1. Pick a source clip
Open Add sound, in the Video section of the sidebar, and choose a clip you already generated, browsing by folder, or import any video file from disk.
2. Describe the sound
Type a short prompt for the ambience you want, in English, for example "footsteps on gravel, light wind" or "rain on a window, distant traffic". Leave it blank and tendre.AI still generates a track matching what is visible in the footage.
3. Keep it music and voice free, or do not
The negative prompt defaults to "music, voice, speech" so the result stays an ambience bed rather than a soundtrack with singing or dialogue. Clear or edit that field if you want to push past the default.
4. Generate and preview
Press Add sound and preview the result against the picture. MMAudio is tuned for short clips, best around 5 seconds and still solid at 10. Each generation covers the whole clip in one pass, so regenerate with a different prompt for another take.
To layer dialogue on top, see Character voices. For sound tuned to a specific action rather than general ambience, see Sound effects.