audio

Audio playback, editing and analysys

// js
audio('raw.wav').trim(-30).normalize('podcast').fade(0.3, 0.5).save('clean.mp3')

# cli
audio raw.wav trim -30db normalize podcast fade 0.3s -0.5s save clean.mp3

Any Format — fast wasm codecs, no ffmpeg. In browsers, formats beyond the bundled set fall back to native WebAudio decode.
Non-destructive — virtual edits, infinite undo, instant clone.
Stream-first — playback/encode during decode, realtime editing.
Paged — no 2Gb memory limit, open 10Gb+ files.
Analysis — loudness, spectrum, beats, pitch, chords, key.
Modular – pluggable ops, tree-shakable.
CLI — playback, batch processing, scripting, unix pipes, tab completion.
Cross-platform — browsers, node, deno, bun.

Start Recipes API CLI FAQ Plugins Architecture

Start

Node

npm i audio

import audio from 'audio'
let a = audio('voice.mp3')
a.trim().normalize('podcast').fade(0.3, 0.5)
await a.save('clean.mp3')

Browser

<script type="module">
  import audio from './dist/audio.min.js'
  let a = audio('./song.mp3')
  a.trim().normalize().fade(0.5, 2)
  a.clip({ at: 60, duration: 30 }).play()   // play the chorus
</script>

Codecs load on demand via import() — map them with an import map or your bundler.

Only the root audio import ships a prebuilt browser bundle (dist/audio.js/dist/audio.min.js). Subpath imports (audio/core.js, audio/fn/gain.js, ...) resolve to source ES modules — fine for Node or bundled browser builds, but a bundler is required to use them directly in a browser.

Import map example

<script type="importmap">
{
  "imports": {
    "@audio/decode-mp3": "https://esm.sh/@audio/decode-mp3",
    "@audio/decode-wav": "https://esm.sh/@audio/decode-wav",
    "@audio/decode-flac": "https://esm.sh/@audio/decode-flac",
    "@audio/decode-opus": "https://esm.sh/@audio/decode-opus",
    "@audio/decode-vorbis": "https://esm.sh/@audio/decode-vorbis",
    "@audio/decode-aac": "https://esm.sh/@audio/decode-aac",
    "@audio/decode-qoa": "https://esm.sh/@audio/decode-qoa",
    "@audio/decode-aiff": "https://esm.sh/@audio/decode-aiff",
    "@audio/decode-caf": "https://esm.sh/@audio/decode-caf",
    "@audio/decode-webm": "https://esm.sh/@audio/decode-webm",
    "@audio/decode-amr": "https://esm.sh/@audio/decode-amr",
    "@audio/decode-wma": "https://esm.sh/@audio/decode-wma",
    "@audio/encode-wav": "https://esm.sh/@audio/encode-wav",
    "@audio/encode-mp3": "https://esm.sh/@audio/encode-mp3",
    "@audio/encode-flac": "https://esm.sh/@audio/encode-flac",
    "@audio/encode-opus": "https://esm.sh/@audio/encode-opus",
    "@audio/encode-ogg": "https://esm.sh/@audio/encode-ogg",
    "@audio/encode-aiff": "https://esm.sh/@audio/encode-aiff"
  }
}
</script>

CLI

npm i -g audio
audio voice.wav trim normalize podcast fade 0.3s -0.5s save clean.mp3

Recipes

Clean up

// master a raw take
let a = audio('raw-take.wav')
a.trim(-30).normalize('podcast').fade(0.3, 0.5)
await a.save('clean.wav')

// full restoration chain via ecosystem plugins (see API › Plugins)
a.gate(-45).dehum().deesser().compressor({ threshold: -18 }).limiter({ ceiling: -1 })

// cut 2:00–2:15, smooth the splice
a.remove({ at: 120, duration: 15 }).fade(0.1, { at: 120 })

// find clipped blocks
let clips = await a.stat('clipping')

Compose

// podcast montage
let ep = audio([intro, interview.trim().normalize('podcast'), outro], { crossfade: 0.5 })
await ep.save('episode.mp3')

// voiceover over music
music.gain(-12).mix(voice, { at: 2 })

// ringtone: the chorus + fades
audio('song.mp3').crop({ at: 45, duration: 30 }).fade(0.5, 2).normalize().save('ringtone.mp3')

// split an audiobook into chapters
let [ch1, ch2, ch3] = audio('audiobook.mp3').split(1800, 3600)

// glitch: stutter + reverse
let v = a.clip({ at: 1, duration: 0.25 })
audio([v, v, v, v]).reverse({ at: 0.25, duration: 0.25 })

Analyze

// waveform bars — and progressively, as it decodes
let [mins, peaks] = await a.stat(['min', 'max'], { bins: canvas.width })
a.on('data', ({ delta }) => appendBars(delta.max[0], delta.min[0]))

// features for ML
let mfcc = await a.stat('cepstrum', { bins: 13 })
let [loud, rms] = await a.stat(['loudness', 'rms'])

// notes, chords, key
let notes = await a.stat('notes')    // [{time, duration, freq, midi, note, clarity}]
let chords = await a.stat('chords')  // [{time, duration, label, root, quality, confidence}]
let key = await a.stat('key')        // {tonic, mode, label, confidence}

Record & generate

// mic take
let a = audio()
a.record()
// …later
a.stop()
a.trim().normalize()

// tone — any t => sample function
let tone = audio.from(t => Math.sin(440 * Math.PI * 2 * t), { duration: 2 })

// sonify data
let s = audio.from(t => Math.sin((200 + data[t / 0.2 | 0]) * Math.PI * 2 * t) * 0.5, { duration: data.length * 0.2 })

Automate

Any numeric op param accepts a t => value function — the engine samples it during render (sample-accurate for gain/pan, ~3ms steps elsewhere). A breakpoint curve {t, v} does the same and stays serializable (survives toJSON() and the worker boundary):

a.gain(t => -12 * (0.5 + 0.5 * Math.cos(t * Math.PI * 4)))  // 2Hz tremolo in dB
a.lowpass(t => 400 + 4000 * t)                              // filter sweep
a.pan({ t: [0, 2, 4], v: [-1, 1, -1] })                     // curve: L→R→L over 4s
music.ducker({ key: voice })                                // sidechain (plugin)

Stream & persist

// stream to network — encode/playback during decode
for await (let chunk of audio('2hour-mix.flac').highpass(40)) socket.send(chunk[0].buffer)

// serialize edits, restore later
let json = JSON.stringify(a)     // { source, edits, ... }
let b = audio(JSON.parse(json))  // re-decode + replay edits

API

Create

audio(source, opts?) – decode from file, URL, or bytes. Returns instantly — decodes in background.
audio.from(source, opts?) – wrap existing PCM, AudioBuffer, silence, or function. Sync, no I/O.

let a = audio('voice.mp3')                // file path
let b = audio('https://cdn.ex/track.mp3') // URL
let c = audio(inputEl.files[0])           // Blob, File, Response, ArrayBuffer
let d = audio()                           // empty, ready for .push() or .record()
let e = audio([intro, body, outro])       // concat (virtual, no copy)
let f = audio([a, b, c], { crossfade: 2 })  // concat with 2s crossfade
// opts: { sampleRate, channels, crossfade, curve, storage: 'memory' | 'persistent' | 'auto' }

await a    // await for decode — if you need .duration, full stats etc

let a = audio.from([left, right])                 // Float32Array[] channels
let b = audio.from(3, { channels: 2 })           // 3s silence
let c = audio.from(t => Math.sin(440*TAU*t), { duration: 2 })  // generator
let d = audio.from(audioBuffer)                   // Web Audio AudioBuffer
let e = audio.from(int16arr, { format: 'int16' }) // typed array + format

Properties

Format, playback and state — media-element semantics where they apply.

.duration – total seconds (reflects edits).
.channels – channel count.
.sampleRate – sample rate.
.length – total samples per channel.
.currentTime – playback position in seconds (smooth interpolation during playback).
.playing – true during playback.
.paused – true when paused.
.volume – playback volume, 0..1 linear. Settable.
.muted – mute gate, independent of volume. Settable.
.loop – loop playback on/off. Settable.
.playbackRate – live playback speed, 0.0625..16. Settable during playback — ramps smoothly (tape-style varispeed, ~50ms), no clicks. Playback-only; use .speed() to bake.
.ended – true when playback ended naturally (not via stop).
.seeking – true during a seek operation.
.played – promise, resolves when playback starts.
.recording – true during mic recording.
.ready – promise, resolves when fully decoded.
.source – original source reference.
.pages – Float32Array page store.
.stats – per-block stats (peak, rms, etc.).
.edits – edit list (non-destructive ops).
.version – increments on each edit.

Structure

Non-destructive time/channel rearrangement. All support {at, duration, channel}.

.trim(threshold?) – strip leading/trailing silence (dB, default auto).
.shrink(gap?, threshold?) – compress silent pauses to a target gap (seconds, default 0.3) throughout, or within {at, duration}. shrink(0) removes silence entirely. _{≡ FFmpeg silenceremove, Audacity truncate-silence}
.crop({at, duration}) – keep range, discard rest.
.remove({at, duration}) – cut range, close gap.
.insert(source, {at}) – insert audio or silence (number of seconds) at position.
.clip({at, duration}) – zero-copy range reference (an excerpt — unrelated to stat('clipping'), which detects over-0dBFS distortion).
.split(...offsets) – zero-copy split at timestamps.
.pad(before, after?) – silence at edges (seconds).
.repeat(n) – repeat n times.
.reverse({at?, duration?}) – reverse audio or range.
.speed(rate) – playback speed (affects both pitch and duration).
.stretch(factor) – time stretch, preserves pitch. Phase-locked vocoder. Factor accepts a t => f function or {t, v} curve over source time — sliding stretch (continuous tempo envelope): duration becomes ∫factor dt, pitch stays put. Ranged via {at, duration}.
.pitch(semitones) – pitch shift, preserves duration. Positive = higher.
.remix(channels) – channel count: number or array map ([1, 0] swaps L/R). No {at, duration} — channel count can't change mid-timeline.

a.trim(-30)                               // strip silence below -30dB
a.remove({ at: '2m', duration: 15 })      // cut 2:00–2:15, close gap
a.insert(intro, { at: 0 })               // prepend; .insert(3) appends 3s silence
let [pt1, pt2] = a.split('30m')          // zero-copy views
let hook = a.clip({ at: 60, duration: 30 })  // zero-copy excerpt
a.stretch(0.9)                            // slow 10%, preserve pitch
a.pitch(-2)                               // down 2 semitones, preserve tempo
a.remix([0, 0])                           // L→both; .remix(1) for mono

Process

Amplitude, mixing, normalization. All support {at, duration, channel} ranges.

.gain(dB, opts?) – volume. Number, range, or t => dB function. { unit: 'linear' } for multiplier.
.fade(in, out?, curve?) – fade in/out. Curves: 'linear' 'exp' 'log' 'cos'. Adjustable via opts: {start, end} gain levels (0..1 — fade between arbitrary levels, e.g. a duck), {mid} — position of the half-amplitude point within the fade (skews the curve), {at} — anywhere in the timeline. _{≡ Audacity adjustable-fade}
.normalize(target?) – remove DC offset, clamp, and normalize loudness. LUFS presets follow EBU R128 / ITU-R BS.1770-4 (equivalent to FFmpeg loudnorm).
- 'podcast' – -16 LUFS, -1 dBTP.
- 'streaming' – -14 LUFS.
- 'broadcast' – -23 LUFS.
- -3 – custom dB target (peak mode).
- no arg – peak 0dBFS.
- { mode: 'rms' } – RMS normalization. Also 'peak', 'lufs'.
- { ceiling: -1 } – true peak limiter in dB.
- { dc: false } – skip DC removal.
.mix(source, opts?) – overlay another audio (additive).
.crossfade(source, duration?, curve?) – crossfade into another audio. Default 0.5s 'cos' (complementary amplitude, best for similar material); 'equal' for the equal-power law (constant loudness across unrelated material, e.g. two songs). _{≡ FFmpeg acrossfade}
.pan(value, opts?) – stereo balance (−1 left, 0 center, 1 right). Accepts function.
.write(data, {at?}) – overwrite samples with raw PCM.
.transform(fn) – inline processor: (input, output, ctx) => void. Not serialized.

a.gain(-3)                                // reduce 3dB
a.gain(6, { at: 10, duration: 5 })       // boost range
a.gain(t => -12 * Math.cos(t * TAU))     // automate over time
a.fade(0.5, -2, 'exp')                    // 0.5s in, 2s exp fade-out
a.normalize('podcast')                    // -16 LUFS; also 'streaming', 'broadcast'
a.mix(voice, { at: 2 })                  // overlay at 2s
a.crossfade(next, 2)                      // 2s crossfade into next
a.crossfade(next, 0.5, 'linear')          // linear crossfade
a.pan(-0.3, { at: 10, duration: 5 })      // pan left for range

Filter

Biquad filters, chainable. All support {at, duration} ranges.

.highpass(freq), .lowpass(freq) – pass filter.
.bandpass(freq, Q?), .notch(freq, Q?) – band-pass / notch.
.allpass(freq, Q?) – all-pass (phase shift, unity magnitude).
.lowshelf(freq, dB), .highshelf(freq, dB) – shelf EQ.
.eq(freq, gain, Q?) – parametric EQ.
.filter(type, ...params) – generic dispatch.

a.highpass(80).lowshelf(200, -3)          // rumble + mud
a.eq(3000, 2, 1.5).highshelf(8000, 3)    // presence + air
a.notch(50)                               // remove hum
a.allpass(1000)                           // phase shift at 1kHz
a.filter(customFn, { cutoff: 2000 })     // custom filter function

Effect

Audio effects and transformations.

.vocals(mode?) – stereo vocal isolation/removal via mid/side cancellation. 'isolate' (default) keeps center, 'remove' keeps sides. _{≡ SoX oops}
.dither(bits?, {shape?}) – TPDF dithering for bit-depth reduction (default 16-bit). shape:true enables 2nd-order noise shaping — pushes quantization noise above ~Nyquist/2 (audibly quieter at given bit depth).
.crossfeed(freq?, level?) – headphone crossfeed for improved stereo imaging. Default: 700 Hz cutoff, 0.3 level. _{≡ SoX earwax, bs2b}
.resample(rate, {type?}) – sample rate conversion. Non-destructive, chainable, undoable. Upsampling defaults to fast linear interpolation; downsampling defaults to an anti-aliased 32-tap windowed sinc. Use type:'sinc' to force sinc quality, or type:'linear' to force the fastest interpolation.
.crossover(...freqs) – band-splitting crossover: N split points → N+1 bands × channels, band-major order. Linkwitz-Riley 4th order, allpass-aligned — summing bands reconstructs the input flat. _{≡ FFmpeg acrossover}

a.vocals()                                // isolate center-panned vocals
a.vocals('remove')                        // remove vocals (karaoke)
a.dither(16)                              // TPDF dither to 16-bit
a.dither(16, {shape: true})               // noise-shaped (audibly quieter)
a.crossfeed()                             // headphone crossfeed
a.resample(48000)                         // resample to 48kHz (linear)
a.resample(96000, {type: 'sinc'})         // high-quality windowed-sinc
a.resample(22050).gain(-3).save('lo.wav') // chain with other ops

I/O

Read PCM, encode, push. Format inferred from extension.

await .read(opts?) – rendered PCM. { format, channel } to convert.
await .save(path, opts?) – encode + write. { at, duration } for sub-range.
await .encode(format?, opts?) – encode to Uint8Array.
.clone() – deep copy, independent edits, shared pages.
.push(data, format?) – feed PCM into pushable instance. .stop() to finalize.

let pcm = await a.read()                              // Float32Array[]
let raw = await a.read({ format: 'int16', channel: 0 })
for await (let block of a) send(block)                 // async-iterable over blocks
await a.save('out.mp3')                                // format from extension
let bytes = await a.encode('flac')                     // Uint8Array
let b = a.clone()                                      // independent copy, shared pages

let src = audio()                                      // pushable source
src.push(buf, 'int16')                                 // feed PCM
src.stop()                                             // finalize

Playback / Recording

Live playback with dB volume, seeking, looping, live meter.

.play(opts?) – start playback. { at, duration, volume, rate, loop }. .played promise resolves when output starts. a.playbackRate is live — set it mid-playback for smooth tape-style speed ramping.
.pause(), .resume(), .seek(t), .stop() – playback control.
.meter(what, cb?) – live stats during playback. what is a stat name, array of names, or opts. Returns a probe { value, stop() }. Listener-gated (zero cost when nothing subscribes).
.record(opts?) – mic recording. { deviceId, sampleRate, channels }.

a.play({ at: 30, duration: 10 })          // play 30s–40s
await a.played                             // wait for output to start
a.volume = 0.5; a.loop = true             // live adjustments
a.muted = true                             // mute without changing volume
a.pause(); a.seek(60); a.resume()         // jump to 1:00
a.stop()                                  // end playback or recording

let mic = audio()
mic.record({ sampleRate: 16000, channels: 1 })
mic.stop()

.meter(what, cb?) — polymorphic first arg: string → single stat, array → keyed object, opts object → full config. Channel semantics mirror a.stat(): omitted → scalar avg, channel: n → that channel, channel: [0, 1] → per-channel array. Omit cb for pull-style access via probe.value.

a.meter('rms', v => draw(v))                                       // scalar avg across channels
a.meter(['rms', 'peak'], v => draw(v))                             // { rms, peak }
a.meter({ type: 'rms', channel: [0, 1] }, v => draw(v))            // [L, R]
a.meter({ type: 'spectrum', bins: 64, smoothing: 0.15 }, drawFFT)  // Float32Array of mel bins
a.meter({}, ({ delta, offset }) => draw(delta))                    // no type → all block stats

let m = a.meter({ type: 'rms' })                                   // pull form
requestAnimationFrame(function tick() { draw(m.value); requestAnimationFrame(tick) })
m.stop()                                                           // release

Opts: type (stat name, array, or omit for all), channel (n, [n, m], or omit), smoothing (one-pole EMA τ in seconds), hold (peak-hold decay τ in seconds), bins / fMin / fMax (when type: 'spectrum'). Any registered stat works (rms, peak, ms, min, max, dc, clipping, spectrum, or user-registered via audio.stat(...)).

Analysis

await .stat(name, opts?) — without bins returns scalar, with bins returns Float32Array. Array of names returns array of results. Sub-ranges via {at, duration}, per-channel via {channel}.

'db' – peak amplitude in dBFS.
'rms' – RMS amplitude (linear).
'peak' – max absolute amplitude, max(|min|, |max|) (linear, dBFS via 20·log10).
'loudness' – integrated LUFS (ITU-R BS.1770).
'dc' – DC offset.
'clipping' – clipped samples (scalar: timestamps, binned: counts).
'silence' – silent ranges as {at, duration}.
'crest' – crest factor in dB (peak/RMS ratio). Sine ≈ 3dB, square ≈ 0dB.
'centroid' – spectral centroid in Hz (brightness). Higher = brighter.
'flatness' – spectral flatness 0–1. 0 = tonal, 1 = noise.
'correlation' – inter-channel (L/R) phase correlation, −1 to +1. Mono returns 1.
'max', 'min' – peak envelope per bin — use together for waveform rendering.
'spectrum' – mel-frequency spectrum in dB (A-weighted).
'cepstrum' – MFCCs.
'bpm' – tempo in BPM.
'beats' – beat timestamps as Float64Array (seconds).
'onsets' – onset timestamps as Float64Array (seconds).
'notes' – pitch events: [{time, duration, freq, midi, note, clarity}] (YIN).
'chords' – chord sequence: [{time, duration, label, root, quality, confidence}] (NNLS chroma + Viterbi).
'key' – musical key: {tonic, mode, label, confidence} (Krumhansl-Schmuckler).

For BPM/beats/onsets, opts: { minBpm, maxBpm, delta, frameSize, hopSize }. Use a.detect(opts) to get { bpm, confidence, beats, onsets } in one pass. For notes, opts: { frameSize, hopSize, threshold, minClarity }. For chords/key, opts: { frameSize, hopSize, method } ('nnls' or 'pcp').

let loud = await a.stat('loudness')                       // LUFS
let [db, clips] = await a.stat(['db', 'clipping'])        // multiple at once
let spec = await a.stat('spectrum', { bins: 128 })        // frequency bins
let [min, max] = await a.stat(['min', 'max'], { bins: 800 }) // peak envelope for canvas rendering
await a.stat('rms', { channel: 0 })                       // left only → number
await a.stat('rms', { channel: [0, 1] })                  // per-channel → [n, n]
let gaps = await a.stat('silence', { threshold: -40 })    // [{at, duration}, ...]
let bpm = await a.stat('bpm')                             // 120.5
let beats = await a.stat('beats')                         // Float64Array [0, 0.5, 1, ...]
let { bpm, confidence, beats, onsets } = await a.detect() // full pipeline, one pass
let notes = await a.stat('notes')                         // [{time, duration, freq, midi, note: 'A4', clarity}]
let chords = await a.stat('chords')                       // [{time, duration, label: 'Am', confidence}]
let k = await a.stat('key')                               // {label: 'C', mode: 'major', confidence}

Utility

Events, lifecycle, undo/redo, serialization.

.on(event, fn) / .off(event?, fn?) – subscribe / unsubscribe.
- 'data' – pages decoded/pushed. Payload: { delta, offset, sampleRate, channels }.
- 'change' – any edit or undo.
- 'metadata' – stream header decoded. Payload: { sampleRate, channels }.
- 'timeupdate' – playback position. Payload: currentTime.
- 'play' – playback started or resumed.
- 'pause' – playback paused.
- 'volumechange' – volume or muted changed.
- 'ended' – playback finished (not on loop).
- 'progress' – during save/encode. Payload: { offset, total } in seconds.
.dispose() – release resources. Supports using for auto-dispose.
.undo(n?) – undo last edit(s). Returns edit for redo via .run().
.run(...edits) – apply edits as arrays ['type', opts?]. Batch or replay.

Edits use [type, opts] shape, where opts is params (value, freq, etc.) plus range keys (at, duration, channel).

a.run(
  ['gain', { value: -3, at: 10, duration: 5 }],
  ['crop', { at: 1, duration: 2 }],
  ['fade', { in: 1, curve: 'exp' }],
  ['insert', { source: ref, at: 2 }],
  ['gain', { value: -3 }],
)

let saved = JSON.stringify([
  ['gain', { value: -3 }],
  ['crop', { at: 1, duration: 2 }],
])
a.run(...JSON.parse(saved))

a.on('data', ({ delta }) => draw(delta))  // decode progress
a.on('timeupdate', t => ui.update(t))     // playback position

a.undo()                                  // undo last edit
b.run(...a.edits)                         // replay onto another file
JSON.stringify(a); audio(json)            // serialize / restore

Plugins

One mechanism extends everything — ops, stats, codecs. Built-ins register through the same interface. audio.use a package, or define your own with audio.op / audio.stat. See Plugin Tutorial.

audio.use(...plugins) – register plugins: a factory following the @audio contract, a stat { stat, compute }, a codec { codec, test?, decode?, encode? }, a function receiving audio, or a registry name (dynamic import, returns a promise — npm i the package it points at; catalog in Ecosystem).
audio.op(name, fn) – register op. Shorthand for { process: fn }. Full descriptor: { params, process, plan, resolve }.
audio.op(name) – query descriptor. audio.op() – all ops.
audio.stat(name, descriptor) – register stat. Shorthand (chs, ctx) => [...] or { block, reduce, query }.

Contract factories plug in as ops: params get engine automation, curves and click-free ramps; declared tail, latency, streaming: false and sidechain buses are handled by the engine; the CLI resolves registry names and synthesizes --help from param metadata.

import { compressor } from '@audio/dynamics-compressor/audio'
audio.use(compressor)                       // bring-your-own factory
await audio.use('freeverb', 'declick')      // or by registry name

a.freeverb({ room: 0.8 })                   // tail composes automatically
a.declick()                                 // batch plugins run whole-render
music.ducker({ key: voice })                // sidechain via the key option

Stat plugins land on a.stat(name); option values that are audio instances (e.g. similarity's ref) pre-render to PCM:

await audio.use('truepeak', 'structure')
await a.stat('truepeak')                    // −0.4 dBTP (inter-sample, BS.1770)
await a.stat('similarity', { ref: b })

Codec plugins — { codec: fmt, test?(bytes), decode?(bytes), encode?(opts) } — extend what audio() can open (header sniffed via test where magic-byte detection draws a blank) and what save()/encode() can write. Every @audio/decode-* / @audio/encode-* package ships its half as an audio.js manifest — halves merge by format name; the bundled umbrellas keep precedence for formats they already serve (streaming decode stays streaming), so codec plugins matter for standalone hosts and formats beyond the bundled set.

Custom ops and stats are plain descriptors — chainable and queryable like built-ins:

// op: params declares named args → ctx.bits; process receives (input, output, ctx) per 1024-sample block
audio.op('crush', { params: ['bits'], process: (input, output, ctx) => {
  let steps = 2 ** (ctx.bits ?? 8)
  for (let c = 0; c < input.length; c++)
    for (let i = 0; i < input[c].length; i++)
      output[c][i] = Math.round(input[c][i] * steps) / steps
}})

// stat: block function collects per-block, reduce enables scalar queries across blocks
audio.stat('peak', {
  block: (chs) => chs.map(ch => { let m = 0; for (let s of ch) m = Math.max(m, Math.abs(s)); return m }),
  reduce: (blockValues, from, to) => { let m = 0; for (let i = from; i < to; i++) m = Math.max(m, blockValues[i]); return m },
})

a.crush(4)                    // chainable like built-in ops
a.stat('peak')                // → scalar from reduce
a.stat('peak', { bins: 100 }) // → binned array

Plugins also run without the engine: audio/batch hosts one over a whole signal, audio/stream over live chunks — same param semantics (defaults, automation functions, smoothing), no plan or context.

import { toBatch, toStream } from 'audio/batch'
const compress = toBatch(compressor, { sampleRate: 44100 })
const out = compress(samples, { params: { threshold: -24 } })

Note-event instruments (voice, poly) take a notes list — the host compiles it to contract §events slots:

await audio.use('poly')
audio(4).poly({ notes: [{ time: 0, midi: 60, duration: 1 }, { time: 0, midi: 64, duration: 1 }] })

Worker

The whole engine off the main thread — one import, same call shape; the main bundle holds a few-KB facade. See architecture.

import audioWorker from 'audio/worker'
let a = audioWorker('track.mp3')            // decode/edits/stats/encode in a Worker
a.gain(-3).fade(0.5)
let [mins, maxs] = await a.stat(['min','max'], { bins: 640 })  // transferred, zero-copy
a.play()                                    // AudioWorklet (no SharedArrayBuffer) / @audio/speaker
a.playbackRate = 1.5                        // live, smooth-ramped — parity with the local player

Once audio/worker is imported, the main entry dispatches too: audio('track.mp3', { worker: true }) — same call shape, worker-hosted.

Custom worker entry (extra codecs, plugins): import '@audio/decode-aac'; import 'audio/worker' — the file self-hosts in worker scope; pass it via { worker }. Boundary notes: clip()/split()/clone() return promises of facades; op errors surface on 'error' (or await a.run([type, opts])); function params don't cross — use curves {t, v}.

CLI

npm i -g audio

audio [source] [transforms...] [sink] [options]

A pipeline: a source produces audio, transforms reshape it, a sink consumes it. The default sink is stat — printing an overview.

# sources
FILE         path, URL, or glob  ('*.wav' for batch)
-            stdin (or omit when piping)
record       capture from microphone

# transforms (chained left-to-right)
gain         fade        trim        normalize   crop
clip         remove      reverse     repeat      pad
speed        stretch     pitch       insert      mix
crossfade    remix       pan         split       resample
highpass     lowpass     eq          lowshelf    highshelf
notch        bandpass    allpass     vocals      dither
crossfeed    shrink      crossover

# sinks (terminate the chain — at most one)
stat [NAMES...]    print analysis (default)
play [loop]        open player UI
save PATH          encode and write (or `-` for stdout)

# options
-f --force         overwrite existing output
--format FMT       override output format
--macro FILE       apply edits from JSON
--cue FILE         split at cue-sheet tracks (with split)
--verbose          show progress
--help, -h         help (or per-op: `audio gain --help`)

# compatibility shortcuts
-p ⇔ play     -l ⇔ play loop     -o PATH ⇔ save PATH

Playback

␣ pause · ←/→ seek ±10s · ⇧←/⇧→ seek ±60s · ↑/↓ volume · l loop · s save as · q quit

# play full song
audio song.mp3 play

# play fragment
audio song.mp3 10s..15s play

# play and loop a hook
audio song.mp3 30s..45s play loop

# play with effects applied live (streamable ops)
audio song.mp3 normalize broadcast highpass 80hz play

Edit

# clean up
audio raw-take.wav trim -30db normalize podcast fade 0.3s -0.5s save clean.wav

# scope a range (applies to whole chain)
audio in.wav 1s..10s gain -3db save out.wav

# range on a single op
audio in.wav gain -3db 1s..10s save out.wav

# filter chain
audio in.mp3 highpass 80hz lowshelf 200hz -3db save out.wav

# concat
audio intro.mp3 + content.wav + outro.mp3 trim normalize fade 0.5s -2s save ep.mp3

# crossfade into next
audio track1.mp3 crossfade track2.mp3 2s save mixed.wav

# voiceover
audio bg.mp3 gain -12db mix narration.wav 2s save mixed.wav

# split
audio audiobook.mp3 split 30m 60m save 'chapter-{i}.mp3'
audio album.wav split --cue album.cue save '{i} - {title}.mp3'   # cue-sheet tracks, tagged

# record
audio record 30s save voice.wav

Analysis

# overview (default sink)
audio speech.wav

# range overview — `audio FILE 0..10s` ⇔ `audio FILE stat 0..10s`
audio speech.wav 0..10s

# specific stats
audio speech.wav stat loudness rms

# tempo / beat grid / onsets
audio track.mp3 stat bpm
audio track.mp3 stat beats onsets

# pitch / chords / key
audio song.mp3 stat notes
audio song.mp3 stat chords
audio song.mp3 stat key

# spectrum / cepstrum with bin count
audio speech.wav stat spectrum 128
audio speech.wav stat cepstrum 13

# stat after transforms (transforms apply, then stat)
audio speech.wav gain -3db stat db

Batch

audio '*.wav' trim normalize podcast save '{name}.clean.{ext}'
audio '*.wav' gain -3db save '{name}.out.{ext}'

Stdin/stdout

cat in.wav | audio gain -3db save -      > out.wav
curl -s https://ex.com/speech.mp3 | audio normalize save clean.wav
ffmpeg -i video.mp4 -f wav - | audio trim normalize podcast save - > voice.wav

Tab completion

eval "$(audio --completions zsh)"       # add to ~/.zshrc
eval "$(audio --completions bash)"      # add to ~/.bashrc
audio --completions fish | source       # fish

FAQ

What formats are supported?: Decode: WAV, MP3, FLAC, OGG Vorbis, Opus, AAC, AIFF, CAF, WebM, AMR, WMA, QOA via decode. Encode: WAV, MP3, FLAC, Opus, OGG, AIFF via encode. Codecs are WASM-based, lazy-loaded on first use.
Does it need ffmpeg or native addons?: No, pure JS + WASM. For CLI, you can install globally: npm i -g audio.
How big is the bundle?: ~20K gzipped core. Codecs load on demand via import(), so unused formats aren't fetched.
How does it handle large files?: Audio is stored in fixed-size pages. In the browser, cold pages can evict to OPFS when memory exceeds budget — auto-sized from navigator.storage.estimate() (quota/4, 64MB..2GB), overridable via {budget}. Stats stay resident (~7 MB for 2h stereo).
Are edits destructive?: No. a.gain(-3).trim() pushes entries to an edit list — source pages aren't touched. Edits replay on read() / save() / for await.
Can I use it in the browser?: Yes, same API. See Browser for bundle options and import maps.
Does it need the full file before I can work with it?: No. Playback, edits, and structural ops (crop, repeat, pad, insert, etc.) all stream incrementally during decode — output begins before the file finishes loading. The edit plan recompiles as data arrives, tracking a safe output boundary per op. Only ops that depend on total length (open-end reverse, negative at) wait for full decode.
TypeScript?: Yes, ships with audio.d.ts.
How is this different from SoX?: SoX is a C command-line tool — powerful but native-only, no browser, no programmatic API, no streaming edits, no undo. audio runs in Node and the browser with the same API, edits are non-destructive and lazy (nothing is rendered until you read/save), and it streams during decode. Several SoX effects are implemented (allpass, dither, crossfeed/earwax, vocals/oops, resample). Remaining effects (reverb, compressor, noise reduction, chorus, flanger, phaser) are planned.
How is this different from Audacity?: Audacity is a GUI desktop app. audio is a library and CLI — designed for scripting, automation, pipelines, and embedding in apps. Audacity is destructive (edits mutate samples); audio is non-destructive (edits are a plan replayed on read). Audacity can't run in the browser or be npm installed into your project.
How is this different from ffmpeg?: ffmpeg is a video-first tool that also handles audio. It's a C binary — no JS API, no browser, no streaming edits. audio is audio-first: dB, Hz, LUFS are native units. Edits are non-destructive, playback streams during decode, and the whole thing is ~20K gzipped with codecs loading on demand.
How is this different from Web Audio API?: Web Audio API is a real-time audio graph for playback and synthesis — not for editing files. No undo, no save-to-file, no CLI, no Node (without polyfills). audio is for working on audio files: load, edit, analyze, save. For Web Audio API in Node, see web-audio-api.
How is this different from Tone.js / Howler.js?: Tone.js is a Web Audio synthesis framework — great for making music in real-time, not for editing files. Howler.js is a playback library — load and play, no editing or analysis. audio is a complete audio workstation: decode, edit, analyze, encode, play, record, CLI.

Ecosystem

Op plugins (audio.plugins registry, name → package — npm i it, then await audio.use('name')):

dynamics compressor · limiter · gate · expander · deesser · ducker · compand · softclip · leveler · transient-shaper · multiband · fet · opto · varimu · vca — denoise dehum · specsub · wiener · omlsa · dereverb · deplosive · dewind · declick · declip · decrackle · debreath — effects delay · chorus · flanger · phaser · tremolo · vibrato · autowah · wah · bitcrusher · distortion · exciter · ringmod · freqshift · multitap · pingpong · slew · noiseshaper · lofi · graindelay · stutter · subbass · sbr — reverb freeverb · schroeder · plate · fdn · spring · shimmer — filter biquad · moog · korg35 · diode · oberheim · resonator · spectral-tilt · variable · comb · dcblocker · emphasis · deemphasis · derivative · integral — eq geq · tilt · baxandall · dyneq — spatial widener · haas · panner · autopan · midside · microshift · surround — shift pitch-shift · vocoder · formant-shift · paulstretch — color tape · transistor · waveshaper · multisat · amp · cabinet · defeedback — generate osc · noise · chirp · pluck · risset · rhythm · sfx · kick · cymbal · snare · adsr · voice · poly — more yin · tube · isolate · tune

Stat plugins (land on a.stat(name)):

loudness truepeak · lra · replaygain · dr · speech-contrast · sounds — spectral rolloff · spread · slope · flux · contrast · ltas · zcr — mir structure · tempogram · melody · downbeat · fingerprint · drums · multif0 · transcribe · similarity · coversong · chroma · tonnetz

Beyond the registry — kernels whose inputs aren't scalar params ship as plain packages for direct import: @audio/reverb-convolution (impulse response), @audio/eq-fir (response curve), @audio/eq-crossover (SOS designer), @audio/tune-midi (guide notes), @audio/denoise-repair (regions), @audio/synth-dtmf (digit string), @audio/synth-wavetable (tables), per-band forms of multiband/dyneq/multisat, and the @audio/measure, @audio/sinusoidal, @audio/voice tool/substrate families.

Foundations:

decode – codec decoding (13+ formats)
encode – codec encoding
filter – filters (weighting, EQ, auditory)
speaker – audio output
mic – audio input
pitch – pitch, chord, key analysis
audio-type – format detection
pcm-convert – PCM format conversion

MIT · ॐ

Name		Name	Last commit message	Last commit date
Latest commit History 557 Commits
.github		.github
.work		.work
bench		bench
bin		bin
docs		docs
fn		fn
test		test
.esbuild.js		.esbuild.js
.gitignore		.gitignore
README.md		README.md
audio.d.ts		audio.d.ts
audio.js		audio.js
batch.js		batch.js
cache.js		cache.js
core.js		core.js
license.md		license.md
logo.svg		logo.svg
package-lock.json		package-lock.json
package.json		package.json
plan.js		plan.js
player.gif		player.gif
player.tape		player.tape
preview.svg		preview.svg
stats.js		stats.js
worker.d.ts		worker.d.ts
worker.js		worker.js

Uh oh!

Uh oh!

Folders and files

Latest commit

History

Repository files navigation

audio

Start Recipes API CLI FAQ Plugins Architecture

Start

Node

Browser

CLI

Recipes

Clean up

Compose

Analyze

Record & generate

Automate

Stream & persist

API

Create

Properties

Structure

Process

Filter

Effect

I/O

Playback / Recording

Analysis

Meta

Utility

Plugins

Worker

CLI

Playback

Edit

Analysis

Batch

Stdin/stdout

Tab completion

FAQ

Ecosystem

About

Topics

Resources

License

Uh oh!

Stars

Watchers

Forks

Releases 5

Sponsor this project

Uh oh!

Packages 0

Uh oh!

Uh oh!

Contributors

Uh oh!

Languages

Packages