Skip to content
Get started
GENERATION API CALLS
Audio Generation

Scenario

This page is auto-generated from model configurations. Last updated: 2026-11-16.

This reference lists all available Scenario audio generation models and their parameters. Use these parameter names when calling the Generation API.


Trim an audio file to a precise time range with sample-accurate cutting.

Model ID: model_scenario-audio-cut

Capabilities: audio2audio

LLM Markdown: https://app.scenario.com/api/models/model_scenario-audio-cut/markdown

Parameter Type Required Default Min Max Allowed Values Description
audio file Yes - - - - The audio file to trim. Only the part between Start Time and End Time is kept.
startTime number No 0 0 - - Start time in seconds. Default is 0 (beginning of audio).
endTime number No - 0 - - End time in seconds. If not specified, cuts to the end of the audio.
outputFormat string No - - - ``, mp3, wav, ogg, m4a Output audio format. If not specified, the original format is preserved.

Pull the audio out of any video and save it as a standalone MP3, WAV, or AAC file.

Model ID: model_scenario-audio-extract

Capabilities: video2audio

LLM Markdown: https://app.scenario.com/api/models/model_scenario-audio-extract/markdown

Parameter Type Required Default Min Max Allowed Values Description
video file Yes - - - - The video file to extract audio from. The existing audio track is pulled out as-is — no generation or modification applied.
outputFormat string No mp3 - - mp3, wav, aac The file format for the extracted audio. MP3 is widely compatible; WAV is lossless; AAC offers a good balance of quality and file size.
normalizeLoudness boolean No false - - - Adjusts the volume of the extracted audio to a consistent, broadcast-safe level. Useful if the original video is too quiet or uneven. Enabling this re-encodes the audio.

Split an audio file at one or more timestamps into ordered segments.

Model ID: model_scenario-audio-split

Capabilities: audio2audio

LLM Markdown: https://app.scenario.com/api/models/model_scenario-audio-split/markdown

Parameter Type Required Default Min Max Allowed Values Description
audio file Yes - - - - The audio file to split into segments at the cut points you set.
cutPoints number_array No `` 0 - - Sorted timestamps in seconds at which to split the audio. N cut points produce N+1 segments. Leave empty for one segment containing the full audio.
outputFormat string No - - - ``, mp3, wav, ogg, m4a Output format for all segments. If not specified, the source format is preserved.
strict boolean No false - - - If enabled, reject malformed cut points (unsorted, duplicates, out of range, or empty). If disabled, cut points are normalized automatically.

Split a song into its vocals, drums, bass and other instruments, each as its own audio file.

Model ID: model_scenario-audio-stems

Capabilities: audio2audio

LLM Markdown: https://app.scenario.com/api/models/model_scenario-audio-stems/markdown

Parameter Type Required Default Min Max Allowed Values Description
audio file Yes - - - - The song to split, up to 10 minutes.
stems string_array No vocals,drums,bass,other - - vocals, drums, bass, other The parts to return, one audio file each. Guitars, keyboards and everything that is not vocals, drums or bass end up in Other.
outputFormat string No mp3 - - mp3, wav, flac Format of the stem files.

Music and sound effects composed by Claude Opus 5.5 as Strudel code, from a prompt, an image, a reference audio, a video or your own samples. Returns the audio and its editable program.

Model ID: model_scenario-audio-studio

Capabilities: txt2audio, audio2audio, video2audio

LLM Markdown: https://app.scenario.com/api/models/model_scenario-audio-studio/markdown

Parameter Type Required Default Min Max Allowed Values Description
prompt string No - - - - What to make: style, mood, instruments, tempo, key and structure, like “warm lo-fi hip-hop at 80 BPM with a dusty piano”, or the sound effect, like “a retro coin pickup”. Optional when you add an image, an audio, a video, samples or Strudel code.
audioType string No music - - music, sfx Music, or sound effects. Without a duration, music lasts 30 s and sound effects 3 s.
image file No - - - - Optional. An image to score: the music follows its mood and setting.
audio file No - - - - Optional. A recording to accompany, match or rebuild: the result follows its tempo and key, starting on its first beat. Not with a video.
video file No - - - - Optional. A video to score: hits land on its cuts, and it comes back with the new soundtrack. Not with a reference audio.
samples file_array No - - - - Optional. Up to 8 of your own sounds for the result to play: one-shots (a drum hit, a single note), loops or phrases. The program plays them as sample1, sample2 and so on, in this order: a note is tuned and becomes melodies and chords, a loop plays at its own tempo, and loops and phrases can be chopped. Each keeps its first 60 s.
transcribeAudio boolean No false - - - With a reference audio: split it into stems, transcribe their notes and play them on new instruments in the style of the prompt. Needs Rights Confirmed.
rightsConfirmed boolean No false - - - Confirms that the music rebuilt from the reference audio is yours to use, or in the public domain. Required by Rebuild From Notes.
mixWithInput boolean No false - - - Keep the reference audio, or the video’s own sound, under the result.
returnVideo boolean No true - - - With a video: also return it with the new soundtrack.
duration number No - 0.5 180 - Seconds, up to 180. Leave empty to match the video or the reference audio (up to 180 s), else 30 s of music or 3 s of sound effects. Longer results cost more.
thinkingLevel string No high - - low, medium, high, xhigh How much Claude thinks before each draft. Higher levels write richer arrangements and cost more.
maxIterations number No 5 1 10 - How many times Claude may render, listen to and revise the piece, the final render included. More drafts polish the result and cost more.
code string No - - - - Optional. A Strudel program, like the code a previous run returned: rendered as is without a prompt or media, revised by Claude with them.
outputFormat string No mp3 - - mp3, wav, flac Format of the delivered audio.