A browser multitrack recorder is a fair weekend build: Web Audio for recording and playback, clips on a timeline, per-track volume and pan, a few effects and a WAV mixdown. That isn't what Soundtrap sells. The plans pay for tens of thousands of licensed loops and one-shots, sampled instruments, vocal tuning and mastering styles, and a studio where several people record into one project together. A prompt can't conjure a licensed library, and low-latency shared editing is its own project. If you only want somewhere to layer ideas, the free tools below already do that.
Before you buildSoundtrap Free is a real starting point, with unlimited projects and a smaller library. BandLab, Audacity and LMMS are free as well, so try them before building anything.
Build me a local multitrack sketchpad to replace the solo-studio side of Soundtrap. Requirements: - Node.js 22 + TypeScript: a Vite and React front end using `tone` (Tone.js) for the transport and the Web Audio API for the rest, plus a small Fastify server that stores projects. `npm start` serves http://127.0.0.1:4870 only. - Storage: each project is a folder ./projects/<name>/ with project.json (tempo, time signature, tracks, clips) and audio/*.wav, so a project can be copied, zipped and backed up as plain files. - Tracks: audio tracks plus one drum-machine track. Each has volume, pan, mute, solo, a colour and an insert chain of up to four effects (EQ, compressor, reverb, delay) built from Web Audio nodes, with the reverb impulse response generated in code. - Recording: pick an input from a list, capture with an `AudioWorklet` to 32-bit float WAV, play a metronome with a count-in, and keep a latency offset in milliseconds with a "calibrate" button that records a click played through the speakers and measures the delay. - Timeline: clips snap to a bar and beat grid; drag to move, trim, loop and split at the playhead; zoom, and undo or redo of the last 100 edits. Waveform peaks are cached in ./projects/<name>/peaks/. - Drums: a 16-step pattern editor for kick, snare, hat, clap and tom, all synthesized in code so no samples ship with the app, rendered to audio clips on demand. - Loops: import .wav, .mp3 and .flac files from LOOP_DIRS in .env into a browser panel, audition them at the project tempo, drag them onto a track, and time-stretch with `soundtouchjs`. - Export: a 24-bit WAV mixdown through `OfflineAudioContext` and one WAV stem per track, with an optional peak-normalize step. - Single user. No accounts, no telemetry, local-first: nothing is uploaded. - Out of scope: the loop library itself, real-time collaboration and chat, vocal tuning, mastering presets, VST/AU plugins, MIDI keyboards, mobile. - README: setup, a headphones warning, how latency calibration works, and the browser differences (Safari handles AudioWorklet and microphone permission differently).
Opens your agent with the prompt filled in, you press enter. Or copy it and paste it into an empty folder.
The first prompt builds the core. Run these one at a time in the same folder to grow it into your own Soundtrap.
Add an importer for Soundtrap exports to the multitrack sketchpad in this folder. Requirements: - A command `sketch import-stems <folder> --bpm <n>` that takes the per-track WAV files exported from a Soundtrap project and creates a new project with one track per file. - Place every clip at bar 1, set the project tempo from the --bpm flag, and name tracks after the file names. - Warn when the files differ in length or sample rate, and resample to the project rate. - Update the README with how to export tracks from Soundtrap.
Add take comping to the multitrack sketchpad in this folder. Requirements: - Recording over an existing region keeps each pass as a separate take in a take lane under the track instead of overwriting it. - Drag across a take to select a range and it becomes the active audio for that range; the track shows the resulting comp and plays it with a 5 ms crossfade at each join. - Flatten a comp to a single clip, and keep the unused takes in the project folder. - Update the README.
Add a vocal-cleanup chain to the multitrack sketchpad in this folder. Requirements: - A "clean vocal" insert preset: high-pass at 80 Hz, a noise gate with adjustable threshold, a de-esser around 6 to 8 kHz and a gentle compressor, all from Web Audio nodes. - Show gain reduction meters for the gate and the compressor while the track plays. - Offline-render the chain when exporting so the WAV matches what I heard. - Update the README with the default values and what each knob does.
Beginners get a ready kit in a browser tab: drums, bass, synths and loops that already fit together, and a way to invite a friend into the same project from another city. The free plan holds back the larger library, the vocal tools and WAV downloads, so anyone who wants those pays by the month.
Estimate · Runs on your own machine with no running cost. The build is token usage on your agent plan and is larger than for most entries because audio timing needs several rounds of testing.
| Plan | Monthly | Annual, per month | What you get |
|---|---|---|---|
| Free | $0 | $0 | Unlimited projects, 7,690+ loops, 1,480+ one-shots, 450+ instruments and sounds, 20+ vocal presets, essential effects and 150k+ sound effects from freesound.org. |
| Sound Starter | $11.99 | $9.59 | 32,560+ loops, 7,430+ one-shots, 1,190+ instruments and sounds, 40+ effects, 30+ vocal presets, vocal tuning, automation and 3 mastering styles. |
| Music Production | $16.99 | $13.59 | 39,480+ loops, 9,660+ one-shots, 1,800+ instruments and sounds, 50+ effects, 8 mastering styles, priority mixing and high-quality (WAV) downloads. |
| Vocals & Songwriting | $16.99 | $13.59 | 32,560+ loops, 150+ vocal presets, realtime vocal tuning, effect presets for vocals, 8 mastering styles, 8 hours of interactive transcripts, WAV downloads and publishing a podcast to Spotify. |
| Production & Vocals | $20.99 | $16.59 | Everything in both plans above with 9,800+ one-shots, plus VST/AU plugin support and a one-month Auto-Tune 2026 trial. |
Free gives unlimited projects with a smaller library: 7,690+ loops, 450+ instruments and sounds, 20+ vocal presets and essential effects.
Not really. Beginners get a ready kit in a browser tab: drums, bass, synths and loops that already fit together, and a way to invite a friend into the same project from another city. The free plan holds back the larger library, the vocal tools and WAV downloads, so anyone who wants those pays by the month. The prompt on this page builds the closest honest substitute.
About $11.99 a month on the Sound Starter plan, $143.88 a year (checked 2026-10-19).
The licensed library of tens of thousands of loops, one-shots and instruments; Real-time collaboration with video chat inside the same project; Vocal tuning, vocal presets and the mastering styles; VST/AU plugin support and the Auto-Tune trial on the top plan; The mobile app and publishing a podcast to Spotify. If any of those matter to you, keep paying.
Yes: Audacity (open source), LMMS (open source), Ardour (open source).