m4l-builder

Build Max for Live devices in Python. Write scripts, emit .amxd files straight to your Ableton User Library. No Max GUI required. Everything is version-controllable, scriptable, and reproducible. Zero runtime dependencies -- pure stdlib.

Install

pip install m4l-builder

Quick start

from m4l_builder import AudioEffect, WARM, device_output_path

device = AudioEffect("My Gain", width=150, height=110, theme=WARM)

device.add_panel("bg", [0, 0, 150, 110])
device.add_dial("gain", "Gain", [10, 6, 50, 90],
                min_val=-70.0, max_val=6.0, initial=0.0,
                unitstyle=4, annotation_name="Gain")

device.add_newobj("mul_l", "*~ 1.", numinlets=2, numoutlets=1, outlettype=["signal"])
device.add_newobj("mul_r", "*~ 1.", numinlets=2, numoutlets=1, outlettype=["signal"])
device.add_newobj("db2a", "dbtoa", numinlets=1, numoutlets=1, outlettype=[""])
device.add_newobj("pk", "pack f 20", numinlets=2, numoutlets=1, outlettype=[""])
device.add_newobj("ln", "line~", numinlets=2, numoutlets=2, outlettype=["signal", "bang"])

device.add_line("obj-plugin", 0, "mul_l", 0)   # stereo in
device.add_line("obj-plugin", 1, "mul_r", 0)
device.add_line("mul_l", 0, "obj-plugout", 0)   # stereo out
device.add_line("mul_r", 0, "obj-plugout", 1)
device.add_line("gain", 0, "db2a", 0)           # dial -> dbtoa -> smooth -> gain
device.add_line("db2a", 0, "pk", 0)
device.add_line("pk", 0, "ln", 0)
device.add_line("ln", 0, "mul_l", 1)
device.add_line("ln", 0, "mul_r", 1)

device.build(device_output_path("My Gain"))

Restart Ableton (or refresh the browser) and your device shows up in the User Library.

Key concepts

Device types

Three device types, matching what Ableton expects:

from m4l_builder import AudioEffect, Instrument, MidiEffect

fx    = AudioEffect("FX", 300, 170)      # auto-adds plugin~/plugout~
synth = Instrument("Synth", 400, 200)    # no auto I/O
midi  = MidiEffect("MIDI FX", 200, 100)  # MIDI only

AudioEffect auto-wires stereo plugin~ / plugout~ objects (IDs obj-plugin, obj-plugout). Instruments and MIDI effects leave I/O to you.

DSP blocks

All DSP functions return (boxes, lines) tuples. Compose them with add_dsp:

from m4l_builder import gain_stage, highpass_filter

boxes, lines = gain_stage("gain")
device.add_dsp(boxes, lines)

Filters: highpass, lowpass, bandpass, notch, onepole, shelves, tilt EQ, 3-band crossover, peaking EQ, allpass Dynamics: compressor, limiter, envelope follower, gate/expander, sidechain detect, multiband compressor Delay/Reverb: delay line, feedback delay, reverb network, FDN reverb, convolver Modulation: LFO (4 waveforms), tremolo, transport LFO, morphing LFO MIDI: notein/out, ctlin/out, velocity curve, transpose, arpeggiator, chord, pitch quantize, midi learn Synthesis: wavetable osc, noise, oscillator bank, ADSR, poly voices, grain cloud Spectral: spectral gate, crossover, vocoder, phase vocoder Routing: selector, send/receive (signal + message), matrix mixer, sidechain routing Utility: param smooth, tempo sync, sample and hold, bitcrusher, coll/dict/pattr storage

90+ blocks total. See docs/api.md for the full list with signatures.

Engines (jsui visualizations)

JavaScript generators for Max's jsui object. Each returns an ES5 string for mgraphics/Cairo rendering:

from m4l_builder.engines import filter_curve_js

device.add_jsui("display", [10, 30, 200, 80],
                js_code=filter_curve_js(), numinlets=3)

18 generators available: filter curves, EQ, envelopes, spectrum, waveforms, XY pad, piano roll, step grids, grain clouds, vocoder bands, and more.

Themes

Seven built-in color themes. Pass to the device constructor and all UI elements inherit colors automatically:

from m4l_builder import AudioEffect, MIDNIGHT, WARM, COOL, LIGHT, FOREST, VIOLET, SOLAR

device = AudioEffect("FX", 300, 170, theme=MIDNIGHT)

Build a theme from just an accent color:

from m4l_builder import Theme
my_theme = Theme.from_accent([0.8, 0.3, 0.1, 1.0])

Recipes

Pre-wired DSP combos for common patterns. They add objects to a device and return IDs for further wiring:

from m4l_builder import gain_controlled_stage, dry_wet_stage

ids = gain_controlled_stage(device, "out", [10, 10, 50, 70])
# ids["gain"] is the *~ you wire into your signal chain

Available: gain_controlled_stage, dry_wet_stage, tempo_synced_delay, midi_note_gate.

Layout helpers

Automatic positioning with Row, Column, and Grid context managers:

with device.row(10, 10, spacing=8, height=70) as r:
    r.add_dial("d1", "Param1", width=50)
    r.add_dial("d2", "Param2", width=50)
    r.add_dial("d3", "Param3", width=50)

Subpatchers

Nested patchers for organizing signal chains:

from m4l_builder import Subpatcher

sub = Subpatcher("processor")
sub.add_newobj("in", "inlet~", numinlets=1, numoutlets=1)
sub.add_newobj("out", "outlet~", numinlets=1, numoutlets=1)
sub.add_line("in", 0, "out", 0)
device.add_subpatcher(sub, "proc_box", [30, 100, 80, 20])

Round-trip editing with from_amxd

Read an existing .amxd back into a Device, modify it, write it out:

device = AudioEffect.from_amxd("path/to/effect.amxd")
device.add_dial("new_knob", "NewParam", [10, 10, 50, 70])
device.build("path/to/modified.amxd")

Examples

40+ examples in examples/. A few highlights:

File	What it builds
simple_gain.py	Minimal starter, single gain dial
stereo_delay.py	L/R delay with feedback saturation
simple_compressor.py	Threshold, ratio, attack/release
parametric_eq.py	JSUI custom EQ curve display
poly_synth.py	Polyphonic synth with wavetable
midi_arpeggiator.py	MIDI arpeggiator
from_amxd_demo.py	Round-trip: build, read back, modify

uv run python examples/simple_gain.py           # run one
for f in examples/*.py; do uv run python "$f"; done  # build all

Development

git clone https://github.com/alaarab/m4l-builder.git
cd m4l-builder
uv pip install -e .

# Tests
uv run pytest tests/ -v

# With coverage
uv run pytest tests/ --cov=m4l_builder --cov-report=term-missing

API reference

Full API docs at docs/api.md.

License

MIT