Sample audio
Create an editable instrumental sample with melody, bass and drums.
Start with a musical idea. Shape the notes, build a rhythm, and make the mix your own.
Melody
One bar, sixteen steps. Choose a note or leave a rest.
Bass
One bar, sixteen steps. Choose a note or leave a rest.
Rhythm
Toggle the steps to build your own beat.
Mix
Space & groove
Melody and bass repeat across the arrangement. The pad follows a four-chord progression.
Recipe import & API example
curl -X POST https://socket.help/api/tools/test-media-audio-composer/run \
-H 'Content-Type: application/json' \
--data-binary @request.json
request.json:
{
"input": "{\"version\":1,\"bpm\":78,\"bars\":8,\"root\":60,\"scale\":\"major\",\"voice\":\"keys\",\"seed\":917,\"swing\":0,\"reverb\":0.32,\"delay\":0.14,\"melody\":[60,null,64,null,67,64,null,62,60,null,65,null,67,null,64,null],\"bass\":[36,null,null,43,null,null,36,null,null,null,41,null,43,null,null,null],\"kick\":[true,false,false,false,false,false,false,false,true,false,false,true,false,false,false,false],\"snare\":[false,false,false,false,true,false,false,false,false,false,false,false,true,false,false,false],\"hat\":[true,false,true,false,true,false,true,false,true,false,true,false,true,false,true,false],\"levels\":{\"melody\":0.7,\"bass\":0.55,\"drums\":0.3,\"pad\":0.28}}"
}Synthesized on your device in a background worker. No recording is uploaded. The library presets use original sounds; this composer does not generate speech or imitate artists.
About Sample audio
Create a complete, editable instrumental audio sample with melody, bass, drums and pads. Start from six musical presets or build your own pattern.
When to use it
Make a custom instrumental loop for a prototype, video, game or audio-player test instead of repeating a single test tone.
Choose Paper Lanterns → change the melody and percussion → render and export your arrangement.How to use it
- Start from a preset or import an audio recipe.
- Edit the notes and drum steps, then balance the four tracks.
- Render the arrangement, listen to the waveform preview and export the WAV and recipe.
Build an arrangement, not a single sound
The composer has four independent parts: a lead melody, a bass line, a three-piece drum pattern and a harmonic pad. Melody and bass each have sixteen editable steps. Choose a MIDI note or a rest for each position, then repeat the pattern across one, two, four or eight bars. The pad supplies a four-chord progression underneath the arrangement.
Changing the key transposes existing melody and bass notes together. Harmony changes the pad’s scale; it does not overwrite notes you have chosen manually. The lead can use warm keys, a plucked sound, glass-like bells or a soft synthesizer. These sounds are created mathematically rather than loaded from sampled recordings.
Shape the groove and the mix
Switch individual kick, snare and hi-hat steps on or off to create the rhythm. The mix controls set the level of each part, and a switch can mute a track without removing its pattern. Swing delays alternate steps slightly. Delay adds repeating echoes, while reverb adds a short spatial tail.
A quieter pad and a sparse melody usually leave more room for narration. For a game loop, try a stronger rhythm and a shorter arrangement. Render after each meaningful change so you can hear its effect. Moving a control updates the recipe and clears the old render, preventing a stale audio preview from being mistaken for the current composition.
Export audio and keep an editable source
The WAV download contains 44.1 kHz, 16-bit stereo PCM. Arrangements are limited to twenty-nine seconds and five MiB, so long, slow arrangements need fewer bars. The waveform comes from the actual PCM samples. Playback is manual, with seeking and a downloadable file available after rendering.
Export the JSON recipe when you want to reproduce the result, version it with a project or continue editing later. The same recipe can be passed to the hosted tool as a JSON string in input. Browser rendering runs in a background worker without uploading the composition. Hosted rendering uses the same engine and your normal successful-run allowance.
What this generator supports
This is an instrumental sequencer and synthesizer. It does not turn a text prompt into a song, generate speech, clone a voice or imitate a named artist. The built-in arrangements are original starting material. If you enter a melody from another composition, the export does not check or clear rights to that melody.
Limits & privacy
Original procedural synthesis, with no uploaded recordings, speech generation or remote AI model. 55–180 BPM, one to eight bars, at most 29 seconds. 44.1 kHz, 16-bit stereo PCM WAV. Audio plays only when you choose Play. Browser rendering stays on your device; hosted rendering uses your daily allowance. Library recordings have CC0 terms; custom melodies are supplied by you.
Use this tool from the API reference or connect through the SDK, CLI, SSH, or MCP. Hosted calls share your plan allowances.