Sonilo Video-to-Music

Hand Sonilo a finished video and it composes an original score to the cut — pacing, motion, and emotion matched, with transitions and beat drops aligned to the edit, at exactly the video's length. This is Sonilo's flagship capability. Every track is licensed (music licensed via Shutterstock) and cleared for commercial use on social, brand content, and advertising.

Setup: See the [setup-api-key](../setup-api-key) skill to connect the Sonilo MCP server and authenticate — sonilo login (no key) or SONILOAPIKEY.

⚠️ Cost: every tool below makes an API call that may incur charges. Only call it when the user has actually asked for a generation. Check getaccountservices (see the [account](../account) skill) if you're unsure whether free-trial runs remain.

Transport: MCP or CLI

Pick one at the start of the session and stay on it. Do not mix the two inside a single job, and do not announce the choice.

1. Sonilo MCP tools visible in this session (videotomusic and friends) — use them. This is the preferred path: it needs no shell, and it is the only one that survives a very long generation. If a call fails to authenticate — rather than failing on its inputs — this transport is not usable in this session: go to 2 instead of retrying it.
2. No usable Sonilo MCP tools, but sonilo account exits 0 — use the CLI commands below. Same API, same account, same credential file. Probe with sonilo account, not sonilo whoami: whoami exits 0 even when signed out, so it cannot tell the two states apart.
3. Neither — stop and run the [setup-api-key](../setup-api-key) skill. Do not call api.sonilo.com with curl to work around it; both transports handle uploads, polling and retries that a bare request does not.

Quick Start

MCP tool call (recommended)

video_to_music(
    video_path="~/Desktop/trailer.mp4",
    prompt="Build suspense, then resolve with a warm cinematic finish"
)

Saves the generated file(s) to SONILOMCPBASE_PATH (~/Desktop by default) and returns the saved path(s) as text.

Python (pip install sonilo)

from sonilo import Sonilo

client = Sonilo()  # reads SONILO_API_KEY

score = client.video_to_music.generate(video="trailer.mp4", prompt="Build suspense, then resolve with a warm cinematic finish")
score.save("score.m4a")

# video_to_video_music: get the video back with the music muxed in
video = client.video_to_video_music.generate(video="trailer.mp4", prompt="cinematic, uplifting")
video.save("scored.mp4")

preservespeech=True on videotomusic requires the async path — use generateasync() instead of generate() to get the extra vocals/mux/ducked outputs; see sonilo-python's README for the full pattern.

JavaScript / TypeScript (npm install sonilo)

import { SoniloClient } from "sonilo";

const client = new SoniloClient(); // reads SONILO_API_KEY

const score = await client.videoToMusic.generate({
  video: "./trailer.mp4",
  prompt: "Build suspense, then resolve with a warm cinematic finish",
});

// video_to_video_music: get the video back with the music muxed in
const video = await client.videoToVideoMusic.generate({
  video: "./trailer.mp4",
  prompt: "cinematic, uplifting",
});

preserveSpeech: true on videoToMusic requires the async path — use .submit() + client.tasks.wait() instead of .generate(); see sonilo-js's README for the full pattern.

CLI (npm install -g sonilo-cli or pip install sonilo-cli)

sonilo video-to-music --video trailer.mp4 --prompt "Build suspense, then resolve with a warm cinematic finish" --output score.m4a

--format wav, --preserve-speech, and --isolate-vocals each switch video-to-music to the async submit-and-poll path.

# the muxed video, from the CLI
sonilo video-to-video-music --video trailer.mp4 --prompt "cinematic, uplifting" --output scored.mp4

cURL (raw REST API, no MCP host)

curl -X POST "https://api.sonilo.com/v1/video-to-music" \
  -H "Authorization: Bearer $SONILO_API_KEY" \
  -F "[email protected]" \
  -F "prompt=Build suspense, then resolve with a warm cinematic finish" \
  --output score.m4a

videotomusic also accepts a video_url form field instead of an uploaded file — pass one or the other, never both.

Tools

Parameters

Stems

stems=true on videotomusic additionally returns each generated track split into four separated instrument tracks — drums, bass, vocals, other — free of charge. The full mix is untouched; the stems arrive alongside it in the task result as a stems array next to audio:

"stems": [
  {
    "stream_index": 0,
    "drums":  { "url": "…", "content_type": "audio/mp4", "file_size": 2913044 },
    "bass":   { "url": "…", "content_type": "audio/mp4", "file_size": 2870211 },
    "vocals": { "url": "…", "content_type": "audio/mp4", "file_size": 2794560 },
    "other":  { "url": "…", "content_type": "audio/mp4", "file_size": 3011830 }
  }
]

What matters when you use it:

• It splits the GENERATED music, never the video's own audio. The vocals stem is whatever singing the generated score contains — usually near-silent, since scores are mostly instrumental, and that is correct behavior, not a bug. To get the source speech isolated, use preserve_speech (a different feature) instead.
• Async only on REST. stems=true requires mode=async (a 400 otherwise): you get a 202 + taskid and poll /v1/tasks/{taskid}. The MCP tools are always async, so on the hosted server the param just works.
• Available on every surface (verified 2026-08-17): REST, the hosted MCP server, the local sonilo-mcp package (>= 0.18.0), the SDKs (sonilo npm >= 0.16.0, PyPI >= 0.15.0), and the CLIs (--stems, npm sonilo-cli >= 0.15.0, PyPI sonilo-cli >= 0.14.0).
• Match stems to tracks by stream_index, never by array position. A stream whose separation failed is simply absent, so stems can be shorter than audio.
• stemserror is not a failed generation. When separation failed wholly or partly, or was skipped, the task carries a stemserror string — possibly alongside a partial stems array. The generation itself succeeded and every audio URL is valid: treat missing stems as a missing extra, never as a reason to retry or refund.
• Timing: separation runs after generation finishes — typically another 2–6 min, giving up after 30 min (then stems_error).
• The four stem names are fixed (htdemucs separation): melodic instruments — piano, synths, guitar, strings — land in other.
• Formats: stems normally follow outputformat; trust each stem's contenttype for what was actually delivered.

Prompting

No prompt is required — the model reads the cut. A short structured brief adds what the video can't carry: your intent — genre, the moment that must hit, the sounds that must not appear.

Before a paid call: probe the exact duration and existing audio, respect the 360 s cap (over = 422 reject, never truncated), and get sign-off — failed runs auto-refund, but your own retry is a new charge. Write the brief first, generate once, iterate on the prompt, not on rerolls.

• Full pre-flight (inspect the video, caps, credits, verification): [references/preflight.md](../references/preflight.md)
• Style-prompt craft (audio brief, genre/energy wording, preserve_speech, segmented music): [references/music-prompting.md](../references/music-prompting.md)

Workflow Tips

• Prompting: a style hint (e.g. "cinematic, uplifting") steers the result, but the prompt is optional — Sonilo already reads the footage.
• preservespeech for talking-head or narrated video: if the source has dialogue/voiceover the user wants kept, set preservespeech=true. Behavior differs by tool: on videotomusic you get the music, the isolated speech stem, and a ready-mixed combined file (the mux) — use the mux directly rather than re-mixing yourself. On videotovideo_music there's no separate stem or mux file; it just keeps the source speech audible in the single muxed output video.
• Want the video back with the score baked in? Use videotovideomusic instead of videotomusic — same inputs, but the output is a new .mp4, not just audio. Warn the user that the source audio is dropped by default: if their video has dialogue, narration, or effects they expect to hear, pass keeporiginalsound=true (add ducking=true to keep the voice on top), or preservespeech=true for the isolated speech only. A "the music is there but my voiceover vanished" report is always this.
• Several takes in one go: variants_num=3 returns three distinct directions for one request instead of three re-rolls. It costs 3×, and it is never free-trial covered — say the price before calling.
• User wants the score's instruments as separate files (to remix, re-balance, or drop one)? stems=true on videotomusic — it's free, but async-only and not on every surface yet, and it splits the generated music, never the source audio; see [Stems](#stems).
• Duration: always matched to the source video automatically — don't ask for it.
• Need SFX too? To generate music and sound effects for the same video in one balanced, single-charge call, use [video-to-sound](../video-to-sound) rather than calling this and [video-to-sfx](../video-to-sfx) separately.
• Don't know what it should sound like? Run [video-analysis](../video-analysis) first: one call returns a section plan plus ready-to-use generation prompts read off the footage, which beats guessing a prompt and rerolling. It is a paid call that generates nothing, so use it when the brief is genuinely unclear — not when the user already told you what they want.
• Content restriction: prompts cannot reference specific artists, bands, or copyrighted lyrics.

Recovering a Timed-Out Call

videotomusic (without any of preservespeech/ducking/stems/outputformat="wav") streams its result in one call. Any variant that triggers the backend's async mode — and videotovideomusic, which is always async — can time out on a very long TIMEOUTSECONDS. If it does, the error message includes a taskid; the generation keeps running (and is already charged) on the backend. Call getsfxtask(taskid) — getgenerationtask(taskid) on the hosted server — to retrieve the result once ready; see the [task-recovery](../task-recovery) skill.

Output Files

• videotomusic: .m4a by default (.wav if requested), named from the prompt (slugified) or sonilo-<timestamp>.m4a. Multiple parallel streams get a -<index> suffix.
• videotomusic(preserve_speech=true): also saves <base>-vocals.<ext> (isolated speech) and <base>-mux.<ext> (speech+music combined — the one to actually use).
• videotovideo_music: a single .mp4 with the score muxed in.

Error Handling

Common errors: 401 invalid key, 402 insufficient balance / trial exhausted, 413 file too large, 422 invalid parameters (e.g. video over the 360 s cap), 429 rate limit. See the [account](../account) skill to check trial/usage before a call, and the sonilo-mcp README's error table for exact recovery steps.