ElevenLabs launched Instant Soundtracks on October 15, 2026. You upload a video, the model writes an original score for it, then analyzes the footage for sound-effect cues and places generated effects at those timestamps. It is available in ElevenCreative, the ElevenMusic platform and the Video to Music API. There is no separate price: Eleven Music uses about 900 credits per minute, and music commercial use starts on the $6-a-month Starter plan.
Last checked: October 26, 2026.
How we checked: we have not run Instant Soundtracks itself. Everything about the feature below comes from ElevenLabs' launch post, API reference, changelog and pricing page, and the demo video is ElevenLabs' own. The two audio clips further down were made from text prompts on RaoMusic on October 26, 2026, to show the manual alternative.
What is ElevenLabs Instant Soundtracks?
The launch post describes four changes to how ElevenLabs scores video:
- Synced sound effects. After writing the music, the model scans the footage for effect cues and places each sound on the timeline. ElevenLabs' example is "a whoosh timed to the logo at the end of an ad."
- Mood changes between scenes. The backing track can move from EDM or R&B in a fast scene to acoustic guitar in a slower one.
- Your clip's own audio as a reference. The model can read the existing audio's genre, crescendos, beats per minute or instrumentation and shape the new score to match.
- Tighter short-form timing around cuts, transitions, and on-screen actions.
The launch post says the feature is available in the API from day one. The API changelog added a Video to Music entry dated October 19, 2026.
ElevenLabs' official demo of Instant Soundtracks, embedded from its YouTube channel. It shows a video being scored and sound effects landing on on-screen moments.
How do you use Instant Soundtracks in ElevenCreative?
According to ElevenLabs, the feature lives in the Music Video Editor inside ElevenCreative's music tab:
- Upload your video. The model scores it and adds sound effects at the timestamps it picks.
- Turn on the audio-as-reference control if you want the new score to follow the music already in your clip.
- Review the result. Sound effects sit on their own layers, separate from the music, so you can move a hit, add more effects or swap the music without starting over.
- Export the finished video from ElevenCreative.
Check the plan list on the pricing page before you upload a long project, since credits come from the same monthly pool as ElevenLabs' other products.
What changed in the Video to Music API?
The API call is POST /v1/music/video-to-music. It takes one or more video files and returns a single audio file by default. These are the settings that matter most, taken from the API reference:
| Setting | What it does | Default |
|---|---|---|
Automatic sound effects (add_sfx) | Adds sound effects based on cues in the video | false |
Separate files (separate_tracks) | Returns a ZIP with the music and the effects as separate files | false |
Follow the clip's audio (use_audio_as_reference) | Uses the video's audio for genre, BPM, instrumentation, mood and arc | false |
Music model (model_id) | music_v1, music_v2 or music_v2_5 | music_v1 |
All three switches are off by default, so a plain request returns music without effects, and the model defaults to music_v1. One request takes up to 10 videos, 200 MB and 600 seconds in total, joined in order and scored as one video. 192 kbps MP3 output needs the Creator tier or above, and 44.1 kHz PCM needs Pro or above.
How much does Instant Soundtracks cost?
Neither the launch post nor the pricing page lists a separate price for Instant Soundtracks or Video to Music. What the pricing page does say:
- All products draw from one shared monthly credit pool, and Eleven Music uses about 900 credits per minute.
- The Free plan includes 10,000 credits a month. Starter is $6 a month with 30,000 credits.
- "Music commercial use" appears in the Starter plan's feature list, not the Free plan's.
ElevenLabs does not publish a separate Video to Music rate or an extra charge for add_sfx on its pricing page or in the API reference. Run a short clip first and check how many credits it used before you score a full video.
What did the manual route return?
Without video-aware scoring, you make the two halves separately and place the effect yourself. We rebuilt ElevenLabs' own logo-reveal example this way on RaoMusic, with ElevenLabs' music and sound-effect models.
Prompt: 30-second background track for a product ad: upbeat modern electronic pop, punchy drums, bright synth plucks, builds to a confident ending · Engine: ElevenLabs v2 · Simple mode · Instrumental · Length 30 s · MP3
Listening note: punchy electronic drums, bright synth plucks and a clean build to a decisive final hit, close to the prompt. The synth hook is repetitive and fairly generic, but the track works as is under a 30-second product ad or social clip.
Prompt: fast airy whoosh rising into a soft deep impact, for a logo reveal at the end of an ad · Engine: ElevenLabs sound effects · Length 2 s · MP3
Listening note: a smooth, airy riser that lands on a warm low thud, as asked. The impact is soft rather than punchy, so it suits a clean, minimal logo reveal or title card.
The music cost 20 credits and was ready about 11 seconds after we sent the request; the whoosh cost 5 credits and was done within about 6 seconds (we checked every 5 seconds). You still have to find the frame where the logo lands and line the whoosh up with it in a video editor. That placement is the step Instant Soundtracks does for you.
How does it compare with other ways to score a video?
| ElevenLabs Instant Soundtracks | RaoMusic Video to Music | Text prompt + separate effect | |
|---|---|---|---|
| Input | Up to 10 videos, 200 MB, 10 min total | 1 clip (MP4, MOV, WebM), 100 MB, 241 s | Text only |
| Sound effects | Generated and placed automatically | Not included | Generated separately, placed by you |
| Separate music and effect files | Yes, ZIP via API | Music only | Yes, two files |
| Reference audio | The video's own audio track | Optional reference audio clip you attach | None |
| Price | ElevenLabs credits; no separate rate published | RaoMusic: 20 credits for two versions on Rao v1.0 | RaoMusic: 40 credits per minute of music, 1 credit per second of effect (minimum 5) |
| Matches the clip's length | Not stated in the docs | No, trim or fade it in your editor | You choose the length |
ElevenLabs credits and RaoMusic credits are separate currencies, so the numbers in the price row cannot be compared directly.
RaoMusic's Video to Music page reads the clip's mood and returns two versions, instrumental by default, but it does not add sound effects. For the effect layer, the AI sound effect generator makes short effects from a description, like the whoosh above. Music made on RaoMusic comes with royalty-free commercial usage rights on every plan, subject to its license terms; the commercial license page explains that this is a usage license, not a copyright transfer.
Who is Instant Soundtracks for?
Going by ElevenLabs' description, it suits short ads, product demos and social clips where cuts and on-screen moments should land on a sound, and teams that want the music and the sound effects as separate files so they can mix them themselves. Developers scoring many clips can use the API with add_sfx and separate_tracks. If your video only needs a music bed under narration, a text prompt or a video-to-music tool without effects covers it; our guide to background music for YouTube videos walks through levels and length. For how ElevenLabs Music compares with Suno on full songs, see ElevenLabs vs Suno.
Frequently Asked Questions
ElevenLabs' launch post and API reference describe the feature as music plus sound effects and do not mention a vocal or lyrics option.
The API reference offers a zero retention mode, switched on by setting enable_logging to false, for enterprise customers. It does not say how long uploaded videos are kept otherwise, so check ElevenLabs' privacy terms.
Not yet. RaoMusic runs ElevenLabs Music and ElevenLabs sound effects from text, plus its own Video to Music tool, but it does not include ElevenLabs' video-aware scoring.
Yes, through the API. The Video to Music request takes an optional text description of up to 4,100 characters and up to 10 style tags; without them, the model works from the video alone.
Only if you ask for it. The API has a sign_with_c2pa option that signs the file with C2PA content credentials. It works for MP3 output and is off by default.