Audio I/O for Python,
running in the browser

Record the microphone and play audio back from Python inside the browser — Pyodide, JupyterLite, thebe — straight to and from a NumPy array. No native PortAudio, no server.

Try recording → How it's built →

rec = record(3.0) recorded 3.00 s

rec.samplesfloat32 · (144000, 1) · 48 kHz

Features

01Record Capture mic input to a float32 NumPy array — then play your voice as an instrument. → 02Play Push any NumPy buffer to the speakers. → 03Synth Design a sound in NumPy — pianos, plucked strings, a drum kit — then play it live with your computer keyboard. → 04Real-time A NumPy callback fills each block as the browser plays it — polyphonic synths, FM with live knobs, and a live score you edit while it loops. → 05sounddevice facade Drop-in browseraudio.sounddevice: query_devices, default, single-cell play, a real-time OutputStream, and a rec that guides you to the two-cell flow. → 06Genie mode Play the whole piano with 8 buttons — Piano Genie (Donahue, Simon & Dieleman, 2019) picks the notes. →

Why it exists

Python can't reach the speakers, so it asks the page

In the browser, Python usually runs in a Web Worker with no access to getUserMedia or the Web Audio API — so the usual sounddevice → PortAudio stack can't run. browseraudio does the capture and playback on the page's main thread, through a small anywidget frontend, and ships the samples across to the kernel.

Read the architecture →