Browser & Presets

Walk the browser, load presets and drum kits, and generate presets via the preset-XML backdoor.

Tools on this page
  1. get_browser_tree
  2. get_browser_items_at_path
  3. load_drum_kit
  4. list_device_presets
  5. load_preset
  6. list_wavetable_tables
  7. create_wavetable_preset
  8. create_simpler_instrument
  9. inspect_rack_preset
  10. clone_rack
  11. rename_rack_macros
  12. set_rack_zones
  13. add_rack_macro_mappings

get_browser_tree 

get_browser_tree(category_type: str = 'all', limit: int = 200)

Get the browser category tree from Ableton, flattened to a depth-first node list (each node carries its depth, so the hierarchy is reconstructable).

Parameters:

  • category_type: Type of categories to get (‘all’, ‘instruments’, ‘sounds’, ‘drums’, ‘audio_effects’, ‘midi_effects’)
  • limit: Max nodes to return (default 200). The list is truncated past this to stay within size limits; narrow category_type or raise limit.

Returns JSON {category_type, total_folders, count, returned, truncated, nodes:[{name, path, depth, has_more}, …]}. count is the full total; nodes holds at most limit of them.

get_browser_items_at_path 

get_browser_items_at_path(path: str, limit: int = 200)

Get browser items at a specific path in Ableton’s browser.

Parameters:

  • path: Path in the format “category/folder/subfolder” where category is one of the available browser categories in Ableton
  • limit: Max items to return (default 200). The list is truncated past this to stay within size limits; narrow path or raise limit for more.

Returns JSON {path, count, returned, truncated, items:[…]} (plus any other fields Live returns). count is the full total; items holds at most limit.

load_drum_kit 

load_drum_kit(track_index: int, rack_uri: str, kit_path: str)

Load a drum rack and then load a specific drum kit into it.

Parameters:

  • track_index: The index of the track to load on
  • rack_uri: The URI of the drum rack to load (e.g., ‘Drums/Drum Rack’)
  • kit_path: Path to the drum kit inside the browser (e.g., ‘drums/acoustic/kit1’)

list_device_presets 

list_device_presets(path: str, max_depth: int = 6, limit: int = 100)

Recursively list loadable presets under a browser path.

Use this to discover presets for a device/category without knowing their cryptic URIs, then load one with load_preset (pass a preset’s name or its relative_path — load_preset searches subfolders, so either works).

Parameters:

  • path: Browser path like ‘instruments/Wavetable’ or ‘audio_effects/Reverb’ (same grammar as get_browser_items_at_path; category is one of instruments, sounds, drums, audio_effects, midi_effects).
  • max_depth: How many levels of subfolders to descend (default 6).
  • limit: Max presets to return (default 100). Factory devices can expose hundreds; the result is truncated past this to stay within size limits — narrow path to a subfolder or raise limit for more.

Returns JSON {path, count, returned, truncated, presets:[{name, relative_path, uri}, …]}. count is the full total; presets holds at most limit of them.

load_preset 

load_preset(track_index: int, path: str, name: str = None, track_type: str = 'regular')

Load a preset onto a track by readable name/path (no URI required).

Works for any loadable browser item, including instrument/effect presets (.adv) and rack/group presets (.adg). Loading onto a track that already holds an instrument HOT-SWAPS (replaces) it, matching Live’s behavior.

Parameters:

  • track_index: The index of the track to load onto (ignored for master).
  • path: If name is given, the folder to search (e.g. ‘instruments/Wavetable’) — subfolders are searched recursively. If name is omitted, the full path to a single preset (e.g. ‘instruments/Wavetable/Ambient & Evolving/Dark Swarm.adg’).
  • name: Optional preset name OR relative_path (as shown by list_device_presets) to match under path. Matched exact, then case-insensitive, then unique substring, across name and relative_path. Note factory preset names include their file extension (e.g. ‘Dark Swarm.adg’).
  • track_type: ‘regular’ (default), ‘return’, or ‘master’. Only audio effects are valid on return/master tracks.

Tip: run list_device_presets first, then pass a preset’s name or relative_path here (it resolves nested presets, so either works).

list_wavetable_tables 

list_wavetable_tables()

List the factory Wavetable oscillator tables available by name, plus the unison-mode legend — the inputs for create_wavetable_preset.

The oscillator wavetable selection, unison mode, and unison voice count are NOT in Wavetable’s automatable parameter list, so they can’t be set with set_device_parameter. create_wavetable_preset reaches them by writing a preset; this tool tells you the valid table names to pass it.

Returns JSON {count, tables:[…], unison_modes:{name:int}, voices:[min,max]}.

create_wavetable_preset 

create_wavetable_preset(
    name: str,
    table1: str,
    table2: str = None,
    unison_mode: str = 'classic',
    voices: int = 7,
    amount: float = 0.85,
    osc1_pos: float = None,
    osc2_pos: float = None,
    osc2_on: bool = True,
    track_index: int = None,
)

Generate a Wavetable preset (.adv) in the User Library and optionally load it.

This is the ONLY way to set Wavetable’s oscillator table selection, unison mode, and unison voice count via MCP — those controls are absent from the automatable parameter API. The preset is built by patching Live’s factory default, so everything else stays at default and the automatable params (osc detune/pos, unison amount, filter, envelopes) remain tweakable afterwards with set_device_parameter.

Parameters

  • name: preset name (becomes ‘.adv' under User Library; no slashes).
  • table1: factory wavetable for Osc 1 (see list_wavetable_tables, e.g. ‘Basic Shapes’, ‘Quad Saw’, ‘Logue Saw’).
  • table2: factory wavetable for Osc 2 (defaults to table1).
  • unison_mode: name or int 0..6 — None, Classic (supersaw), Shimmer, Noise, Phase Sync, Position Spread, Random Note. Default ‘classic’.
  • voices: unison voice count 2..8 (default 7).
  • amount: unison amount 0.0..1.0 (default 0.85).
  • osc1_pos / osc2_pos: optional wave position 0.0..1.0 (left at default if omitted).
  • osc2_on: enable Osc 2 (default True, for a 2-osc supersaw).
  • track_index: if given, attempt to load the new preset onto that track.

Loading note: Live only sees a brand-new preset file after it re-indexes the User Library (on launch, or a browser refresh) — so a first-time load may report “written but not yet loadable”. Once the name is indexed, re-running this tool overwrites the same file and loads fresh immediately.

Returns JSON {written, path, settings, loaded, note}.

create_simpler_instrument 

create_simpler_instrument(
    name: str,
    sample_path: str,
    root_key: int = 60,
    track_index: int = None,
)

Create a Simpler preset that loads a chosen audio file, and optionally load it.

The sample a Simpler plays is NOT in Live’s automatable parameter API, so a preset is the only way to set it via MCP. This writes a Simpler .adv into the User Library referencing sample_path (Live relinks + recomputes on load). Everything else about Simpler (filter type, ADSR, LFO, warp via the UI) is automatable afterwards with set_device_parameter.

Parameters

  • name: preset name (becomes ‘.adv' under User Library; no slashes).
  • sample_path: absolute path to a .wav or .aif/.aiff file. Other formats are rejected (their frame count can’t be read, which would yield a silent zero-length sample) — convert to WAV/AIFF first.
  • root_key: MIDI note the sample plays at unpitched (default 60 = C3).
  • track_index: if given, load the new preset onto that track.

Returns JSON {written, path, sample, frames, sample_rate, loaded, note}.

inspect_rack_preset 

inspect_rack_preset(template: str)

Inspect a rack preset (.adg) and report its macros and chain zones.

Reads a rack by path (a factory rack under Core Library/Racks/…) or by name (a factory rack name, or a preset already in your User Library) and returns its structure — the inputs you need for rename_rack_macros and set_rack_zones. Reads only; nothing is written or loaded.

Macros are reported 1-based (matching set_rack_macro); chains 0-based (matching get_rack_chains). key_zone/velocity_zone are null on audio-effect and midi-effect racks (those chains have only a chain_selector_zone).

Parameters:

  • template: a .adg path/glob, OR a factory/User-Library rack name.

Returns JSON {rack_type, num_visible_macros, macros:[…16…], chains:[…]}.

clone_rack 

clone_rack(template: str, name: str)

Copy a rack preset into your User Library as a new working preset.

Racks have no blank/default template, so the rack-preset pipeline starts by cloning an existing rack (factory or User Library). The clone becomes the working preset that rename_rack_macros / set_rack_zones edit in place; load the finished result later with load_preset using the returned browser_path.

The User Library category and device-folder are derived from the rack type (instrument -> Instruments/Instrument Rack, audio_effect -> Audio Effects/ Audio Effect Rack, midi_effect -> MIDI Effects/MIDI Effect Rack, drum -> Drums/Drum Rack).

Parameters:

  • template: source rack — a .adg path/glob, or a factory/User-Library name.
  • name: new preset name (becomes ‘.adg'; no slashes or leading dot).

Returns JSON {written, path, browser_path, summary:{…inspect shape…}}.

rename_rack_macros 

rename_rack_macros(
    name: str,
    macro_names: dict = None,
    macro_colors: dict = None,
    macro_defaults: dict = None,
)

Rename / recolor / set defaults for a User Library rack’s macros, in place.

Operates on a rack preset previously created by clone_rack (the working preset). Only the macros you list are changed; the rest keep their values. Macro display names are NOT settable via the LOM — this preset edit is the only MCP route. Macros are addressed 1-based (matching set_rack_macro).

Parameters:

  • name: the User Library rack preset name (from clone_rack).
  • macro_names: object {macro_number: “Display Name”} (no double-quote).
  • macro_colors: object {macro_number: color_int} (-1..69; -1 = no color).
  • macro_defaults: object {macro_number: default_float} (0..127).

The file is overwritten in place, so it reloads fresh with load_preset without re-indexing. Returns JSON {written, path, macros:[…16…]}.

set_rack_zones 

set_rack_zones(name: str, zones: list)

Set chain key / velocity / chain-selector zones on a User Library rack.

Operates in place on a rack created by clone_rack. Chain zones are not settable via the LOM — this preset edit is the only MCP route. Chains are addressed 0-based (matching get_rack_chains). key and velocity are rejected for audio-effect / midi-effect racks (those chains only have a chain_selector zone).

Each entry in zones is an object: {“chain”: 0, “key”: [0, 60] // shorthand [min,max], OR a full object: “velocity”: {“min”:0,”max”:127,”crossfade_min”:0,”crossfade_max”:127}, “chain_selector”: [0, 63]} Provide only the zones you want to change. Validation enforces 0..127 and min <= crossfade_min <= crossfade_max <= max.

Returns JSON {written, path, chains:[…]}.

add_rack_macro_mappings 

add_rack_macro_mappings(name: str, mappings: list)

Map rack macros to device parameters on a User Library rack, in place.

Macro->parameter mappings are NOT settable via the LOM — this preset edit is the only MCP route. Operates on a rack created by clone_rack (the working preset). Macros are addressed 1-based (matching set_rack_macro); chains 0-based (matching get_rack_chains).

Each entry in mappings is an object: {“macro”: 1, “chain”: 0, “device”: 0, “parameter”: “Gain”}

  • macro: 1..16 macro number.
  • parameter: the parameter’s XML tag name within the device (e.g. “Gain”, “Volume”). Use inspect-style trial or Live’s device to find tag names; an invalid name returns an error listing the device’s available tags.
  • device address — either the flat single-level form:
    • chain: 0-based chain index.
    • device: 0-based device index within that chain. OR, to target a parameter inside a NESTED rack-in-rack, a “path” array of {chain, device} steps that descend one rack level each (the nested rack’s type is auto-detected and may differ from the outer rack’s). The flat form is equivalent to a one-step path. Example (chain 0 -> device 1 is a rack -> its chain 0 -> device 0): {“macro”: 1, “path”: [{“chain”: 0, “device”: 1}, {“chain”: 0, “device”: 0}], “parameter”: “Frequency”} Supplying both “path” and flat chain/device is rejected as ambiguous.
  • range (optional): {“min”: , "max": } to make the macro sweep a SUB-range of the parameter instead of its full range. min/max are in the parameter's NATIVE units (the same units as get_device_parameters' min/max/value — e.g. linear amplitude, Hz, dB), NOT 0..127. Setting min > max INVERTS the mapping (macro-up drives the param down). Omit for the default full-range mapping.

Mappings are default full-range unless range is given. Re-mapping a parameter replaces its existing mapping; several parameters may share one macro. Load the result with load_preset, then verify by moving the macro (set_rack_macro) and reading the device parameter.

Returns JSON {written, path, mappings:[…]}.

ableton-mcp-max - free for noncommercial use; commercial licensing available. Built by Ryan Robson / Robworks Software LLC. Not affiliated with or endorsed by Ableton AG.

This site uses Just the Docs, a documentation theme for Jekyll.