Send a Video, Get a Score: ElevenLabs Opens Video to Music to Developers

ElevenLabs has launched a Video to Music API that generates a music score and sound effects that follow a video’s cuts and action. It’s live now for developers and enterprises on ElevenAPI.

The same technology powers Instant Soundtracks, a new feature in ElevenLabs’ ElevenCreative platform that adds music and sound effects directly to uploaded videos.

How the Video to Music API Works

You send one or more videos, and the API returns a soundtrack timed to the edit. ElevenLabs first offered Video to Music as a tool in its Studio app in 2025, and the API now opens it up to developers building it into their own apps and workflows.

You can guide the result with a text description and up to 10 style tags, like “cinematic” or “upbeat.” The API supports ElevenLabs’ music models, including Music v2.5.

Sound Effects and Separate Tracks

The API can now add sound effects on top of the music, based on what happens on screen. With the separate_tracks setting turned on, music and effects come back as separate files instead of one mixed track.

For editors, separate tracks matter. They let you balance music and effects independently, or replace either one without regenerating the whole soundtrack.

Matching Your Video’s Existing Audio

A setting called use_audio_as_reference analyzes the audio already in your video, including genre, BPM, instrumentation, mood, and arc. It uses that analysis to guide the new score.

That’s useful when you want to replace a temp track or a placeholder with something in the same style without copying it.

Limits and Formats

Each request accepts up to 10 videos, which are combined in the order sent, with a combined limit of 200MB and 10 minutes. ElevenLabs recommends combining videos into one file first when possible, since many separate files can slow down requests.

Output defaults to MP3, with other formats available. Higher-quality MP3 and PCM output require higher ElevenLabs plan tiers, and outputs can optionally be signed with C2PA content credentials.

Competitive Context

The launch came a day after Mirelo released SFX 2.0, which takes a different approach to AI video sound. Mirelo focuses on removing unwanted music and building sound effects around dialogue, covered in Can AI Fix AI Video’s Sound? Mirelo SFX 2.0 Rebuilds Audio Without Regenerating the Clip.

ElevenLabs is focused on the score, with effects layered on top. For a look at its music models, see ElevenLabs Music v2.5 and Suno vs. ElevenMusic.

Traditional royalty-free music libraries like Soundstripe and Artlist remain the go-to for many editors who want curated, human-composed tracks.

The Signal in the Noise

For developers and platforms building video tools, a single API call that returns a timed score and separate effects track is a practical building block. The audio-as-reference option is especially useful for swapping temp music while keeping the feel of a cut.

For individual creators, Instant Soundtracks in ElevenCreative is the easier way in. Whether the music holds up for client or festival work will depend on the project, so it’s worth testing on a rough cut before committing.

The Details

• Product: ElevenLabs Video to Music API
• Availability: Developers and enterprises on ElevenAPI
• Consumer version: Instant Soundtracks in ElevenCreative
• Inputs: Up to 10 videos, 200MB and 10 minutes combined
• Guidance: Text description, up to 10 style tags, optional audio reference
• Models: Music v1, v2, v2.5
• Outputs: MP3 by default; optional separate music and SFX tracks; optional C2PA signing
• Pricing: Exact pricing available on ElevenLabs’ website

Would you use AI to score a rough cut, or do you still reach for a music library? Share your workflow in the comments.

Resources & Reads

• Video to Music API Reference — ElevenLabs
• Video to Music — ElevenLabs
• ElevenLabs Developers announcement on X
• Can AI Fix AI Video’s Sound? Mirelo SFX 2.0
• Suno vs. ElevenMusic