Background music
Add instrumental background music to any take, generated by DemoTake AI or from your own file. It loops under the whole video, matches loudness and dips automatically while someone speaks.
Music makes a silent product demo feel finished, and a long study video possible. DemoTake treats music like every other edit: it sits on its own lane under the video, it never changes your recording, and Undo or edit_revert takes it away again.
- Added “Chill lofi” (2:28), looping under the take
- Loudness matched to −16 LUFS
- Dips 12 dB under your voice
Add music
In the editor, open the inspector's Effects tab. The Music group has two buttons:
- Generate… makes new instrumental music with DemoTake AI. Pick a style (Chill lofi, Ambient, Soft piano, Upbeat product, Focus electronic, Coffee-shop jazz or Cinematic), add a few words about the mood, and choose a full song or a 30-second clip.
- Import Audio… uses your own MP3, M4A, WAV, AIFF or FLAC file. DemoTake copies it into the take, so the take stays self-contained.
Or type it into the command bar: "add calm piano music", "add a 30-second upbeat clip", "remove the music".
Either way the music starts at 0:00, loops to the end of the video with a smooth crossfade at every seam, fades in and out, and is set well under narration (about 11 dB below a normal voice).
How the mix works
| What | What DemoTake does | Change it |
|---|---|---|
| Looping | A short track repeats with an equal-power crossfade, so there's no click or dip at the seam. | Fade in / Fade out |
| Loudness | Every track is measured when it's added (EBU R128 integrated loudness) and matched to −16 LUFS, without letting its peaks go above −1 dBFS. A loud track and a quiet one sit at the same level. | Volume |
| Ducking | When someone speaks, the music dips (12 dB by default), starting a quarter second before the voice and coming back up over almost a second after. Short pauses between sentences keep it down, so it doesn't pump. | Lower under voice, Dip by |
| Mute | Keeps the music in the take but silent in exports. | Mute |
"Speaking" means the recording's own microphone or system audio, any narration track you import, and a voiceover made in DemoTake. DemoTake finds it by listening to the recorded audio, so music under a silent screen recording plays at full level and dips only where there is sound.
"Lower the music under my voice" in the command bar turns ducking on with a deeper 15 dB dip. "Make the music quieter" lowers it by 4 dB. Both show up in Undo like any other edit.
Playlists for long videos
A two-hour video with one song on repeat gets tiring. Ask for several tracks instead (in Generate…, raise Tracks, or tell your agent music_generate {tracks: 8}). DemoTake generates distinct tracks in the same style, crossfades them into one continuous playlist, matches their loudness, and loops the playlist until the video ends, with a final fade-out on the last second of the video.
Before generating several tracks, DemoTake shows what it will use: "8 songs ≈ 19:04 of music, looped about 8×; uses 8 of this month's songs" (or the estimated cost on your own AI provider key).
Allowances and your own key
| Plan | AI music |
|---|---|
| DemoTake Web free take | One 30-second clip |
| Local (one-time purchase) | With your own AI provider key (Settings › AI); nothing hosted |
| Director | 30 full songs a month |
| Team | 60 songs a month per seat, shared by the team |
| Enterprise | 200 songs a month, or what your contract says |
A 30-second clip counts as half a song. Music has its own allowance: generating music never uses up the AI you use for editing. Generations are also limited per hour, so a runaway agent can't spend the month in one go. The Music group always says where you stand, for example "About 23 songs left this month".
With your own AI provider key, DemoTake sends the request straight from your Mac to your AI provider and nothing goes through DemoTake Cloud.
Every generated track keeps its prompt, style and seed in the take, so you or an agent can make it again or make a variation later.
Using AI music in your videos
AI music generated in DemoTake is yours to put in your videos, including commercial product demos. A few things to know:
- It's marked as AI-generated. Every generated track carries an inaudible watermark that identifies it as AI-generated audio. It doesn't change what you hear, and it can't be removed by DemoTake or by editing.
- Tell viewers where platforms ask. YouTube asks creators to disclose realistic AI-generated content, and names AI-generated music as an example. Disclosure doesn't limit a video's reach or its eligibility to earn money.
- It isn't exclusive. The music model can produce similar music for other people, so don't register generated tracks with YouTube Content ID or other fingerprinting services. That could claim other creators' videos, and theirs could claim yours.
- Don't imitate artists. Prompts that ask for a specific artist's voice or copyrighted lyrics are refused.
DemoTake charges no royalties for generated music. Whether a generated track is protected by copyright at all differs by country, so if a track matters commercially (a brand jingle, a product you sell), treat it like any other asset and check with your own counsel. Your own imported music is yours to license as you always have.
For agents
- music_generatedry_run first for playlists
- music_add
- export_render
music_generate {take_id, prompt, style, length, tracks, dry_run}makes and places the music.dry_run: truereturns the estimate and allowance only.music_add {take_id, path}imports your own file (lane: music, voice or sfx);music_add {take_id, duck: true, duck_db: -15}changes the mix.music_remove {take_id}removes it.edit_revertundoes any of these.
The DemoTake Web editor plays and exports the same mix: the same lanes, fades, loudness and ducking, so a take sounds the same on the Mac and in the browser.