LiveSession MCP
Click on "Deploy Server".
Wait a few minutes for the server to deploy. Once ready, it will show a "Started" state.
In the chat, type
@followed by the MCP server name and your instructions, e.g., "@LiveSession MCP start a mizrahi pop song in A minor at 96 BPM"
That's it! The server will respond to your query, and you can continue using it as needed.
Here is a step-by-step guide with screenshots.
LiveSession MCP
An AI production assistant for Ableton Live. Compose, arrange and mix with Claude (or any MCP client), with real measurements instead of guesses — and work with your Live Sets even when Live is closed.
LiveSession MCP is an independent project. It is not affiliated with or endorsed by Ableton AG; Ableton and Live are trademarks of Ableton AG. It started as a fork of ahujasid/ableton-mcp (MIT).
What it does
Compose | Chords by symbol or roman numeral, voice-led (close, open, spread, drop 2, shell, rootless, quartal) · melodies with phrase shapes · basslines that follow slash chords · drum patterns · euclidean rhythms and polymeters · motif development (invert, retrograde, augment, sequence, additive...) · a library of 45 named progressions searchable by genre and mood · genre song starters |
Arrange | Song structure from scenes with locators (clips cut exactly at section ends) · arrangement clips · take lanes · warp markers |
Sound | Your installed plug-ins by role ("a plate reverb", "a warm compressor") from a scanned catalogue · Live's devices with ready-dialled chain recipes · Drum Rack pads · Simpler · presets · plug-in windows |
Mix with measurements | Integrated loudness (LUFS), true peak, loudness range, tonal balance, stereo correlation, key — from your actual bounce · match a commercial reference and get the EQ moves applied · gain staging from Live's meters · sidechain · mix snapshots |
Live's menu | Save, freeze, bounce in place, group, open the export dialog — things Live's API can't do, run through Live's own menu |
Files (Live closed) | Summarise a Live Set · lint it before sharing (missing samples, plug-ins not installed, tracks routed nowhere...) · diff two versions · export MIDI clips · look inside .adg/.adv presets |
Safe by design | Every tool is marked read-only or destructive · batches run as one undo step · the set is saved before deleting actions so Live keeps the previous version · a journal of every destructive call |
148 tools in 10 groups, 6 workflow prompts, server instructions for the model. Load only the groups you
need (LIVESESSION_TOOLSETS=core,compose,mix,listen) — fewer tools means better tool choices.
Related MCP server: AbletonMCP
Requirements
Ableton Live 12 (recommended) or Live 11 — Suite, Standard or Intro
Windows 10/11 or macOS 12+
Python 3.10+ and uv (recommended)
Install
1. Add the server to your MCP client. For Claude Desktop / Claude Code, in the MCP config:
{
"mcpServers": {
"livesession": {
"command": "uvx",
"args": ["--from", "livesession-mcp[all]", "livesession-mcp"]
}
}
}[all] adds listening (loudness analysis) and plug-in scanning. Without extras the core still works.
2. Install the Remote Script into Live (once, and after each upgrade):
uvx --from livesession-mcp livesession-mcp install3. In Live: restart Live, open Settings → Link, Tempo & MIDI, choose AbletonMCP as a Control Surface (Input and Output: None).
4. Check everything:
uvx --from livesession-mcp livesession-mcp doctordoctor checks Python, the installed script version, the connection to Live, the optional extras and
your folders, and tells you how to fix anything that isn't right (--json for scripts).
On macOS, allow your MCP client (Claude, Terminal...) in System Settings → Privacy & Security → Accessibility so menu commands and window pictures work.
Try it
"Start a mizrahi pop song in A minor at 96 BPM."
"Give me a sad progression for the pre-chorus and write it on the piano with rootless voicings."
"Take the melody in slot 1 and make a sequence a step higher in the next slot."
"I exported the mix — how does it compare to this reference? Fix the tonal balance."
"Lint my project folder before I send it to the mixing engineer."
"What changed between today's version and the backup from this morning?"
Prompts ready to use: start_a_song, make_a_beat, mix_and_master, analyze_and_improve,
prepare_for_collaboration, sound_design.
Configuration
livesession-mcp config show
livesession-mcp config set listen_dir "D:/Bounces"
livesession-mcp config set toolsets core,compose,mix,listen
livesession-mcp config set backup_before_destructive falseSetting | Environment variable | Default |
|
|
|
|
| app data folder |
|
|
|
| — |
|
|
|
|
Tool groups: core, compose, arrange, sound, mix, audio, listen, studio, files, advanced.
What Live's API can't do (and what LiveSession does instead)
Export: no API.
prepare_exportsets the range and opens the dialog;listen_latestanalyses the result.Save, freeze, bounce, group: no API — run through Live's own menu. Grouping several tracks needs them selected by mouse first (the API selects one track).
Third-party plug-in parameters: only those exposed in Live (Configure). The catalogue knows every parameter, and
process_audio_with_pluginsrenders audio through any plug-in offline.Arrangement automation and tempo automation: not reachable — use clip automation.
Detaching a groove from a clip: not possible through the API; prefer
swing_clip.Open dialogs in Live (including trial reminders) block the script until closed.
Privacy and security
Nothing leaves your computer: no usage data is collected (see TERMS.md). The Remote Script
listens on 127.0.0.1:9877 only and has no authentication — never forward that port. Your MCP client and
model provider handle your prompts under their own terms. See SECURITY.md.
Development
pip install -e ".[dev]"
pytest
ruff check .
python -m MCP_Server.remote_script_install --sync-bundle # after editing AbletonMCP_Remote_Script/The Live-side studio_ext.py hot-reloads without restarting Live (copy it next to the installed script
and send ext_reload_studio). See CONTRIBUTING.md and CHANGELOG.md.
License
MIT — see LICENSE. Original work © Siddharth Ahuja; LiveSession MCP © Uri Rosen.
Available Tools
148 toolsadd_effectA
Add an effect or instrument by role — "reverb", "eq", "compressor", "deesser", "delay", "saturation", "limiter", "glue", "plate", "hall", "piano", "strings", "bass", "drums"... — using the best installed plug-in and falling back to Live's own device. prefer forces a specific one.
Inputs: role (required), track_index (required), prefer, track_type. Returns: result
add_notes_to_clipB
Add MIDI notes to a clip.
Parameters:
track_index: The index of the track containing the clip
clip_index: The index of the clip slot containing the clip
notes: List of note dictionaries, each with pitch, start_time, duration, velocity, and mute
user_prompt: The original user prompt that led to this tool call (for telemetry)
Inputs: notes (required), clip_index (required), track_index (required), user_prompt. Returns: result
add_warp_markerB
Add a warp marker at beat_time (pinning the audio at sample_time seconds, or where it currently plays).
Inputs: beat_time (required), clip_index (required), track_index (required), arrangement, +1 more. Returns: result
als_diffARead-only
What changed between two versions of a Live Set: tempo, tracks added/removed/renamed, fader, mute, routing and device changes, clips and note counts, locators, scenes, plug-ins. With only old_path: compares the newest copy in the project's Backup folder (old) with that set (new).
Inputs: old_path (required), new_path. Returns: result
als_export_midiARead-only
Export the MIDI clips of a saved Live Set as .mid files (one per clip, set tempo), with Live closed. track filters by track name; out_dir defaults to '/MIDI Export'. Also estimates the key.
Inputs: path (required), track, out_dir. Returns: result
als_lintARead-only
Check a saved Live Set before sharing, mixing or archiving: missing samples, plug-ins missing or not installed here, uncollected files, tracks routed nowhere, soloed or muted tracks, MIDI without an instrument, faders above +6 dB, default names, no locators. Errors first.
Inputs: path (required). Returns: result
als_summaryARead-only
Summarise a saved Live Set without opening Live: tempo, time signature, length, locators, scenes, every track with its devices and clips, returns, plug-ins used. path = an .als file or a project folder (newest set). detail=True includes full device trees and every clip.
Inputs: path (required), detail. Returns: result
apply_chain_recipeCDestructive
Build a ready-dialled chain from Live's own devices (works on any machine, every parameter set): vocal_pop (EQ Eight HPF + presence, Compressor 3:1, de-ess, Glue), piano_ballad (gentle HPF, warm low-mid dip, light compression), strings_pad (HPF, air, wide), cello_solo, bass_warm, drums_soft, master_soft (gentle glue, EQ tilt, limiter at -1 dB).
Inputs: recipe (required), track_index (required), track_type. Returns: result
build_chainC
Build an effect chain by purpose from whatever is installed: vocal, piano, strings, cello, bass, drums, bus, master.
Inputs: kind (required), track_index (required), track_type. Returns: result
build_song_structureA
Lay a song out in the Arrangement from Session scenes, with a named locator per section. sections: [{"name": "Intro", "scene": 0}, {"name": "Verse 1", "scene": 1}, {"name": "Chorus", "scene": 2, "bars": 8}, ...] — each section copies every clip in that scene row to the timeline (repeated to fill "bars" if given, else the longest clip in the row). Tracks without a clip in the row stay silent.
Inputs: sections (required), start_beat, name_scenes. Returns: result
chord_infoBRead-only
Spell chords without writing them: notes of each voicing, e.g. to check a progression.
Inputs: chords (required), style. Returns: result
clear_clip_automationBDestructive
Clear one parameter's automation from a clip, or all of it when parameter is omitted.
Inputs: clip_index (required), track_index (required), parameter, arrangement, +1 more. Returns: result
clear_notes_from_clipADestructive
Remove all MIDI notes from a Session clip.
Writes are additive (add_notes_to_clip only appends), so to truly modify a clip you clear it first, then add the new notes. Use this with get_clip_notes and add_notes_to_clip for a real read -> modify -> write loop: read the notes, edit the list, clear_notes_from_clip, then add_notes_to_clip the edited notes.
Parameters:
track_index: The index of the track containing the clip
clip_index: The index of the clip slot containing the clip
user_prompt: The original user prompt that led to this tool call (for telemetry)
Inputs: clip_index (required), track_index (required), user_prompt. Returns: result
create_arrangement_clipA
Create an empty MIDI clip directly in the Arrangement at beat start. Fill it with
write_chords/write_melody(arrangement=True, clip_index=).
Inputs: start (required), length (required), track_index (required). Returns: result
create_audio_clipA
Create a new audio clip in an audio track's clip slot by importing a file.
Requires Ableton Live 12.0.5 or newer — the underlying ClipSlot.create_audio_clip Live API was introduced in 12.0.5 and is not available in earlier 12.0.x releases.
Parameters:
track_index: The index of the audio track to create the clip in
clip_index: The index of the clip slot to create the clip in
path: Absolute path to a supported audio file (e.g. a .wav). The target track must be an audio track and the clip slot must be empty.
user_prompt: The original user prompt that led to this tool call (for telemetry)
Inputs: path (required), clip_index (required), track_index (required), user_prompt. Returns: result
create_audio_trackA
Create a new audio track in the Ableton session.
Use this for recorded or imported audio (samples, stems, vocals). For MIDI instruments use create_midi_track instead.
Parameters:
index: The index to insert the track at (-1 = end of list)
user_prompt: The original user prompt that led to this tool call (for telemetry)
Inputs: index, user_prompt. Returns: result
create_clipB
Create a new MIDI clip in the specified track and clip slot.
Parameters:
track_index: The index of the track to create the clip in
clip_index: The index of the clip slot to create the clip in
length: The length of the clip in beats (default: 4.0)
user_prompt: The original user prompt that led to this tool call (for telemetry)
Inputs: clip_index (required), track_index (required), length, user_prompt. Returns: result
create_locatorA
Create a named locator (cue point) in the Arrangement at a beat position.
If a locator already exists at that beat (within ~1e-3 tolerance) it is renamed instead of toggled off. Time is in beats from the start of the arrangement (e.g. 0.0 = start, 16.0 = bar 5 in 4/4).
Parameters:
name: The locator label (e.g. "Chorus", "Verse 1", "Drop")
time: Beat position where the locator should sit
user_prompt: The original user prompt that led to this tool call (for telemetry)
Inputs: name (required), time (required), user_prompt. Returns: result
create_midi_trackB
Create a new MIDI track in the Ableton session.
Parameters:
index: The index to insert the track at (-1 = end of list)
user_prompt: The original user prompt that led to this tool call (for telemetry)
Inputs: index, user_prompt. Returns: result
create_return_trackB
Add a return track (for shared reverb/delay); returns its return index for set_mixer sends.
Inputs: name. Returns: result
create_sceneB
Create a scene (row in Session view) at index (-1 = end), optionally named.
Inputs: name, index. Returns: result
create_song_templateB
Build a Nir Canaan-style singer-songwriter session: Piano, Cello, Strings, Bass, Drums, Lead and Backing Vocal tracks with instruments and chains, plus Vocal Plate / Hall / Delay returns and sensible sends. Uses the best installed plug-ins for each role.
Inputs: tempo. Returns: result
create_take_laneB
Add a take lane to a track (for comping several takes).
Inputs: track_index (required), name. Returns: result
crop_clipCDestructive
Crop a clip to its loop/start-end region.
Inputs: clip_index (required), track_index (required), arrangement. Returns: result
delete_arrangement_clipBDestructive
Delete an Arrangement clip (index as listed by get_arrangement_clips).
Inputs: clip_index (required), track_index (required). Returns: result
delete_clipADestructive
Delete the clip in the given clip slot, freeing it for reuse.
Use this before create_clip when you want to overwrite an existing clip (create_clip itself refuses to write into an occupied slot).
Parameters:
track_index: The index of the track containing the clip
clip_index: The index of the clip slot to clear
user_prompt: The original user prompt that led to this tool call (for telemetry)
Inputs: clip_index (required), track_index (required), user_prompt. Returns: result
delete_deviceBDestructive
Remove a device or plug-in from a track, by index or (fuzzy) name.
Inputs: track_index (required), track_type, device_name, device_index. Returns: result
delete_locatorCDestructive
Delete the locator at a beat.
Inputs: time_beats (required). Returns: result
delete_sceneCDestructive
Delete a scene row.
Inputs: index (required). Returns: result
delete_trackBDestructive
Delete a track or return track. Undo with undo() if it was a mistake.
Inputs: track_index (required), track_type. Returns: result
duplicate_clip_loopC
Double a clip's loop with its notes (Live's 'Duplicate Loop').
Inputs: clip_index (required), track_index (required), arrangement. Returns: result
duplicate_clip_regionB
Copy a time range of a MIDI clip's notes to another beat inside the same clip.
Inputs: clip_index (required), destination (required), track_index (required), region_start (required), +2 more. Returns: result
duplicate_to_arrangementA
Copy a Session-view clip into the Arrangement timeline.
Uses Live's track.duplicate_clip_to_arrangement() API (Live 11 / 12). The clip is placed at destination_time beats from the start of the arrangement on the same track it lives in.
Typical workflow:
create_clip / add_notes_to_clip to build a Session clip
Call duplicate_to_arrangement once per bar/section you need
Call switch_to_arrangement_view to confirm the result in Live
Parameters:
track_index: Index of the track that owns the Session clip
clip_index: Index of the clip slot in that track (Session view)
destination_time: Beat position in the arrangement to place the clip (e.g. 0.0 = start, 8.0 = bar 3 in 4/4)
user_prompt: The original user prompt that led to this tool call (for telemetry)
Inputs: clip_index (required), track_index (required), destination_time (required), user_prompt. Returns: result
duplicate_trackB
Duplicate a track with its clips and devices; the copy lands right after it.
Inputs: track_index (required). Returns: result
expose_plugin_paramsA
Prepare exposing a plug-in's controls to Live (needed when get_device_params shows only 'Device On'). Selects the device, finds its window, lists the exact parameter names to touch (from the catalogue) and any saved GUI layout, and returns the computer-use steps.
Inputs: device_name (required), track_index (required), parameters, track_type. Returns: result
find_pluginsBRead-only
Catalogued plug-ins filtered by role (eq, compressor, limiter, deesser, gate, multiband, reverb, delay, saturation, modulation, mastering, pitch, piano, strings, drums, bass, synth) or name/vendor.
Inputs: role, query. Returns: result
find_progressionsARead-only
The progression library (45 named progressions): search by name, genre or mood — "pop", "sad", "mizrahi", "jazz", "house", "cinematic", "israeli"... Empty query lists them all. Play one with write_named_progression.
Inputs: query. Returns: result
fire_clipC
Start playing a clip.
Parameters:
track_index: The index of the track containing the clip
clip_index: The index of the clip slot containing the clip
user_prompt: The original user prompt that led to this tool call (for telemetry)
Inputs: clip_index (required), track_index (required), user_prompt. Returns: result
fire_sceneC
Launch every clip in a scene row.
Inputs: index (required). Returns: result
focus_live_windowA
Bring an Ableton window (main or plug-in, matched by part of its title) to the front.
Inputs: title (required). Returns: result
freeze_trackADestructive
Freeze a track (Edit > Freeze Tracks) and wait until Live finishes rendering it; flatten=True instead prints it to audio for good (Edit > Bounce Tracks in Place: the track becomes an audio track and its devices are removed — undo restores them). Frees CPU from heavy plug-ins.
Inputs: track_index (required), flatten. Returns: result
gain_stageADestructive
Play the song and watch every track's peak meter for a few seconds, then set faders so each track peaks around target_peak_db (default -6 dBFS) — a clean starting balance before mixing. Tracks that stay silent are left alone. apply=False only reports.
Inputs: apply, seconds, start_beat, target_peak_db, +1 more. Returns: result
generate_melodyBDestructive
Generate a singable melody over a progression: chord tones on strong beats, mostly steps, a phrase shape and breaths. contour: arch, rise, fall, wave, question_answer. density 0..1 (0.3 ballad, 0.7 busy). Change seed for another idea; same seed = same melody.
Inputs: key (required), clip_index (required), track_index (required), low, +9 more. Returns: result
get_arrangement_clipsARead-only
List all clips placed in the Arrangement timeline for a track.
Returns each clip's name, start_time, end_time, length, and type.
Parameters:
track_index: The index of the track to inspect
user_prompt: The original user prompt that led to this tool call (for telemetry)
Inputs: track_index (required), user_prompt. Returns: result
get_browser_items_at_pathBRead-only
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
user_prompt: The original user prompt that led to this tool call (for telemetry)
Inputs: path (required), user_prompt. Returns: result
get_browser_treeARead-only
Get a hierarchical tree of browser categories from Ableton.
Parameters:
category_type: Type of categories to get ('all', 'instruments', 'sounds', 'drums', 'audio_effects', 'midi_effects')
user_prompt: The original user prompt that led to this tool call (for telemetry)
Inputs: user_prompt, category_type. Returns: result
get_clip_notesARead-only
Read all MIDI notes from a Session-view clip.
Returns pitch, start_time, duration, velocity, mute (and extended fields when available).
Parameters:
track_index: Track that owns the clip
clip_index: Session clip slot index
user_prompt: The original user prompt that led to this tool call (for telemetry)
Inputs: clip_index (required), track_index (required), user_prompt. Returns: result
get_device_parametersBRead-only
Read all parameters for a device on a track (name, value, min, max).
Parameters:
track_index: Track that owns the device
device_index: Index into the track's device chain
user_prompt: The original user prompt that led to this tool call (for telemetry)
Inputs: track_index (required), device_index (required), user_prompt. Returns: result
get_device_paramsBRead-only
Read a device's or plug-in's parameters with readable values (e.g. "300 Hz", "-3.0 dB"). query filters by parameter name. For plug-ins exposing only 'Device On', the result carries a hint: use Configure in Live (or plugin_window tools + computer use) to expose more controls.
Inputs: track_index (required), query, track_type, device_name, +1 more. Returns: result
get_drum_padsCRead-only
The pads of the track's Drum Rack (also inside an Instrument Rack): note, name, devices.
Inputs: track_index (required), only_filled. Returns: result
get_groovesBRead-only
Grooves in the Groove Pool and the song's global groove amount.
Inputs: no named parameters. Returns: result
get_locatorsARead-only
All Arrangement locators (name, beat) and the song length.
Inputs: no named parameters. Returns: result
get_metersARead-only
Live peak levels (0.0-1.0, 1.0 = 0 dBFS) of every track, return and the master, read right now. Call while playback runs to spot clipping or quiet tracks without exporting.
Inputs: no named parameters. Returns: result
get_mixerBRead-only
Volume, pan, mute, solo, arm, sends, colour and device names for every track, return and master. Volume is Live's 0.0-1.0 fader value (0.85 = 0 dB); volume_display shows dB.
Inputs: no named parameters. Returns: result
get_plugin_layoutCRead-only
A remembered plug-in GUI layout (control positions as window fractions), if one was saved.
Inputs: plugin (required). Returns: result
get_plugin_presetsBRead-only
List the presets/programs a plug-in exposes to Live, and which one is selected.
Inputs: track_index (required), track_type, device_name, device_index. Returns: result
get_plugin_profileBRead-only
Full catalogue profile of one plug-in: all parameters with ranges and options, presets, roles, and its semantic map (mix/threshold/decay... -> real parameter names).
Inputs: name (required), parameter_filter. Returns: result
get_remote_script_infoARead-only
Report Ableton Remote Script version and capabilities (handshake).
Use this to verify the Live-side bridge matches this MCP server package.
Inputs: user_prompt. Returns: result
get_routingCRead-only
A track's audio/MIDI input and output routing, with every available option.
Inputs: track_index (required), track_type. Returns: result
get_session_infoBRead-only
Get detailed information about the current Ableton session
Parameters:
user_prompt: The original user prompt that led to this tool call (for telemetry)
Inputs: user_prompt. Returns: result
get_session_snapshotARead-only
Capture a full project state snapshot for trajectory recording.
Includes session metadata, every track (mixer, devices, session clips, arrangement clips), optional MIDI notes, and optional device parameters. Used to version musical state S_t → S_{t+1}.
Parameters:
include_notes: Include MIDI note arrays in clips (default True)
include_params: Include device parameter values (default True)
user_prompt: The original user prompt that led to this tool call (for telemetry)
Inputs: user_prompt, include_notes, include_params. Returns: result
get_song_scaleARead-only
The song's key and scale as set in Live 12 (root note, scale name, intervals).
Inputs: no named parameters. Returns: result
get_take_lanesCRead-only
A track's take lanes (comping) and the clips on each.
Inputs: track_index (required). Returns: result
get_track_infoCRead-only
Get detailed information about a specific track in Ableton.
Parameters:
track_index: The index of the track to get information about
user_prompt: The original user prompt that led to this tool call (for telemetry)
Inputs: track_index (required), user_prompt. Returns: result
get_track_treeBRead-only
Every track with its group membership, fold and freeze state.
Inputs: no named parameters. Returns: result
get_warp_markersBRead-only
Warp markers of an audio clip (beat_time in the song, sample_time in seconds in the file).
Inputs: clip_index (required), track_index (required), arrangement. Returns: result
group_tracksA
Group a contiguous range of tracks into a Group Track and optionally name it. One track is grouped straight away. For a range, Live needs the track names selected by mouse (its API selects only one track): click the first track's name and shift-click the last, then call this. Anything other than exactly those tracks getting grouped is undone.
Inputs: last_track (required), first_track (required), name. Returns: result
humanize_clipBDestructive
Loosen an existing MIDI clip: small random timing (beats) and velocity changes.
Inputs: clip_index (required), track_index (required), timing, velocity. Returns: result
import_audio_to_arrangementA
Place an audio file on an audio track's Arrangement timeline at a beat (e.g. a vocal take or a reference).
Inputs: path (required), track_index (required), position. Returns: result
insert_deviceA
Insert one of Live's own devices by name ("EQ Eight", "Compressor", "Utility"...) at a chain position (-1 = end) — faster than a browser search.
Inputs: device (required), track_index (required), position, track_type. Returns: result
inspect_live_apiARead-only
Read-only look at Live's Python API: list the attributes (and simple values) of the object at a dotted path such as 'song', 'song.tracks[0].mixer_device' or 'app.browser'. contains filters names. Use it to check whether Live exposes something before writing a new command.
Inputs: path, contains. Returns: result
learn_plugin_profileA
Record the parameters a loaded plug-in exposes in Live into the catalogue. This is the fallback profile for plug-ins the offline scanner can't load (Keyscape, Pianoteq, LABS, Serum 2, MODO BASS...) and it refreshes what Live exposes for any plug-in after Configure. Then get_plugin_profile, set_control and add_effect can use it.
Inputs: device_name (required), track_index (required), track_type. Returns: result
listen_fileARead-only
Analyse a specific WAV/AIFF/FLAC file, optionally against a reference track (a commercial song in the same style): loudness and tonal-balance differences, level-matched.
Inputs: path (required), reference_path. Returns: result
listen_latestARead-only
Analyse the newest audio file in the listen folder (where bounces are exported): loudness, true peak, tonal balance, stereo, key and likely problems, plus a picture (spectrogram, spectrum, loudness over time). Use after the user (or prepare_export) bounced something from Live.
Inputs: no named parameters. Returns: result
list_exportsBRead-only
Audio files in the listen folder, newest first.
Inputs: no named parameters. Returns: result
list_live_windowsARead-only
List open Ableton Live windows (main window and every floating plug-in window) with their screen position and size — use these coordinates to aim computer-use clicks.
Inputs: no named parameters. Returns: result
list_mix_snapshotsBRead-only
List saved mix snapshots.
Inputs: no named parameters. Returns: result
list_pluginsARead-only
List installed VST3/AU plug-ins Live can load, with vendor folder. Filter with query (name or vendor).
Inputs: query. Returns: result
list_racksBRead-only
List device/rack presets (.adg/.adv) saved in the User Library — e.g. your vocal-chain templates.
Inputs: query. Returns: result
load_browser_uriB
Load any browser item (preset, sample, device, kit) by URI onto a track, return or the master.
Inputs: uri (required), track_type, track_index. Returns: result
load_drum_kitA
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')
user_prompt: The original user prompt that led to this tool call (for telemetry)
Inputs: kit_path (required), rack_uri (required), track_index (required), user_prompt. Returns: result
load_drum_padB
Load a sample or instrument from the browser onto one Drum Rack pad (36 = C1 = first pad). query searches samples/drums/sounds/packs (e.g. "brush snare"); uri loads an exact search_browser hit.
Inputs: note (required), track_index (required), uri, query. Returns: result
load_instrument_or_effectBDestructive
Load an instrument or effect onto a track using its URI.
Parameters:
track_index: The index of the track to load the instrument on
uri: The URI of the instrument or effect to load (e.g., 'query:Synths#Instrument%20Rack:Bass:FileId_5116')
user_prompt: The original user prompt that led to this tool call (for telemetry)
Inputs: uri (required), track_index (required), user_prompt. Returns: result
load_pluginADestructive
Load a plug-in onto a track by (fuzzy) name, e.g. "Pro-Q 4", "valhalla vintage", "Keyscape". Effects append to the end of the chain; instruments replace the track's instrument. track_type: "track" (default), "return" or "master" (track_index is ignored for master)
Inputs: name (required), track_type, track_index. Returns: result
load_rackC
Load a saved rack/preset from the User Library onto a track by (fuzzy) name.
Inputs: name (required), track_type, track_index. Returns: result
load_sampleADestructive
Put a sample into Simpler. query/uri: from Live's browser onto the track (an empty MIDI track gets a new Simpler). path: a .wav/.aif on disk replaces the sample in the track's existing Simpler.
Inputs: track_index (required), uri, path, query. Returns: result
match_referenceA
Compare a bounce with a reference song (level-matched tonal balance, loudness, width) and turn the differences into EQ moves. apply=True puts an EQ Eight named "Ref Match" on the master (or track_type/track_index) with those moves — strength 0..1 of the difference, capped at max_db.
Inputs: mix_path (required), reference_path (required), apply, max_db, +3 more. Returns: result
move_deviceB
Reorder a device in its chain, or move it to another track (target_track_index).
Inputs: track_index (required), new_position (required), track_type, device_name, +3 more. Returns: result
move_warp_markerB
Move the warp marker at beat_time by distance beats (stretches the audio around it).
Inputs: distance (required), beat_time (required), clip_index (required), track_index (required), +1 more. Returns: result
prepare_exportBRead-only
Set the loop/export range and open Live's Export Audio/Video dialog (Live's API has no export call), then return the steps to render into the listen folder for listen_latest / match_reference.
Inputs: start_beat, open_dialog, length_beats. Returns: result
preset_infoARead-only
Look inside an .adg rack or .adv device preset: devices and chains, macro names and values, plug-ins and Max for Live devices it needs, and whether it needs Live Suite.
Inputs: path (required). Returns: result
press_live_shortcutADestructive
Bring Live to the front and press keys there, e.g. "tab" (switch view), "ctrl+shift+r" (export), "shift+tab". Live ignores some synthetic Ctrl shortcuts (Ctrl+S, Ctrl+G): use run_live_menu for those.
Inputs: keys (required), repeat. Returns: result
process_audio_with_pluginsA
Render an audio file through catalogued plug-ins outside Live, to audition settings before committing. chain: [{"plugin": "ValhallaVintageVerb", "params": {"Mix": 25, "Decay": 2.5}}, ...] Values are in the units the plug-in displays (25 = 25 %, 2.5 = 2.5 s, 300 = 300 Hz) or option names; see get_plugin_profile for each parameter's range and units. Then analyse the result with listen_file (optionally against the original).
Inputs: chain (required), input_path (required), output_path (required). Returns: result
quantize_clipCDestructive
Quantize a clip. grid: 1/4, 1/8, 1/8t, 1/8+t, 1/16, 1/16t, 1/16+t, 1/32. amount 0.0-1.0.
Inputs: clip_index (required), track_index (required), grid, amount, +1 more. Returns: result
redoADestructive
Redo N undone actions in Live (same as Ctrl+Y).
Inputs: steps. Returns: result
remove_warp_markerCDestructive
Remove the warp marker at beat_time.
Inputs: beat_time (required), clip_index (required), track_index (required), arrangement. Returns: result
restore_mix_snapshotBDestructive
Bring the mixer (and device parameters) back to a saved snapshot. Tracks are matched by index and name.
Inputs: name (required), include_devices. Returns: result
run_batchADestructive
Run several tools in one call as ONE undo step in Live (one Ctrl+Z reverts them all). steps: [{"tool": "create_midi_track", "args": {"index": -1}}, {"tool": "set_track_name", "args": {"track_index": 3, "name": "Pad"}}, ...]. Results come back in order; with stop_on_error the batch stops at the first failure (earlier steps stay applied — call undo once to revert the batch). Desktop, export and analysis tools can't be batched.
Inputs: steps (required), stop_on_error. Returns: result
save_live_setA
Save the current Live Set (File > Save Live Set). A never-saved set opens Live's Save dialog for a name.
Inputs: no named parameters. Returns: result
save_mix_snapshotA
Save every track's mixer state and exposed device parameter values under a name (A/B compare).
Inputs: name (required). Returns: result
save_plugin_layoutA
Remember where controls sit in a plug-in's window, as fractions of the window (0-1): {"Mix": {"x": 0.07, "y": 0.38}, "Decay": {"x": 0.21, "y": 0.42}}. Reused on any later session.
Inputs: plugin (required), controls (required), window_width (required), window_height (required), +1 more. Returns: result
scan_plugin_catalogA
Profile installed VST3 plug-ins outside Live: every parameter (not just what Live exposes), presets, and a role (eq, compressor, reverb, piano...). Resumes where it stopped; call again while 'remaining' > 0; retry_failed re-tries plug-ins that failed before (e.g. after activating them). Modified binaries (broken signature, R2R imports) are listed as excluded, never loaded.
Inputs: only, force, retry_failed, budget_seconds. Returns: result
screenshot_live_windowARead-only
Screenshot one Ableton window — e.g. a plug-in like "Pro-Q 4" — cropped to just that window. Bring it to the front first (focus_live_window) if something covers it. Image pixel (px, py) maps to screen (x + pxscale, y + pyscale) using the returned x, y, scale.
Inputs: title (required), max_width.
search_browserARead-only
Search Live's whole browser: instruments, audio_effects, midi_effects, sounds, drums, samples, plugins, user_library, packs. Load a result with load_browser_uri.
Inputs: query (required), limit, categories. Returns: result
select_deviceB
Select a device and show Live's device view, so its Configure button and plug-in window toggle are on screen for computer use.
Inputs: track_index (required), track_type, device_name, device_index. Returns: result
session_diffA
See what changed in the open set. mark=True records the current state (do it before an edit); mark=False compares now with the last mark: tracks added/removed/reordered, devices, clips, faders, mute/solo, groups, returns, locators, tempo.
Inputs: mark. Returns: result
set_arrangement_clip_nameA
Set the name of a clip placed in the Arrangement timeline.
Parameters:
track_index: The index of the track containing the clip
clip_index: The index of the clip within track.arrangement_clips, in the same order returned by get_arrangement_clips (i.e. ordered by start_time)
name: The new name for the clip
user_prompt: The original user prompt that led to this tool call (for telemetry)
Inputs: name (required), clip_index (required), track_index (required), user_prompt. Returns: result
set_arrangement_timeA
Move the arrangement playhead to a specific position.
Parameters:
time: Position in beats from the start of the arrangement (e.g. 8.0 = bar 3 in 4/4)
user_prompt: The original user prompt that led to this tool call (for telemetry)
Inputs: time (required), user_prompt. Returns: result
set_audio_clipB
Audio clip settings: gain (0-1), warping, warp_mode (beats, tones, texture, re-pitch, complex, complex pro — use complex pro for vocals/full mixes), pitch in semitones/cents, name.
Inputs: clip_index (required), track_index (required), gain, name, +5 more. Returns: result
set_clip_automationC
Draw automation inside a Session clip as steps.
parameter: with device_index omitted, a mixer control: "volume", "pan" or "send 0"; otherwise a parameter name/index on that device
points: [{"time": beats, "value": raw value}, ...]; each value holds until the next point Example — fade the volume up over 4 bars: points=[{"time":0,"value":0.4},{"time":8,"value":0.6},{"time":16,"value":0.85}]
Inputs: points (required), parameter (required), clip_index (required), track_index (required), +3 more. Returns: result
set_clip_colorC
Set a clip's colour by Live palette index (0-69).
Inputs: clip_index (required), color_index (required), track_index (required), arrangement. Returns: result
set_clip_grooveA
Apply a Groove Pool groove to a clip by name; amount = global groove amount (0-1.3, 0 = no effect). Live's API can't detach a groove again (pick None in the clip's Groove box in Live), so prefer swing_clip for a reversible swing. Load grooves into the pool by double-clicking them in Live's Grooves browser.
Inputs: clip_index (required), track_index (required), amount, groove, +1 more. Returns: result
set_clip_loopC
Set a Session clip's loop region in beats and/or turn looping on/off.
Inputs: clip_index (required), track_index (required), looping, loop_end, +2 more. Returns: result
set_clip_nameA
Set the name of a clip.
Parameters:
track_index: The index of the track containing the clip
clip_index: The index of the clip slot containing the clip
name: The new name for the clip
user_prompt: The original user prompt that led to this tool call (for telemetry)
Inputs: name (required), clip_index (required), track_index (required), user_prompt. Returns: result
set_controlB
Set a musical control on ANY device without knowing its parameter names: control is one of mix, threshold, ratio, attack, release, output, input, decay, predelay, size, feedback, delay_time, cutoff, resonance, high_cut, low_cut, width, ceiling, gain, frequency. device is the device's name on the track. If the plug-in hasn't exposed it to Live, the reply names the exact plug-in parameter and the expose step.
Inputs: device (required), control (required), track_index (required), value, +2 more. Returns: result
set_device_onC
Turn a device on or off (bypass), by index or (fuzzy) name.
Inputs: on (required), track_index (required), track_type, device_name, +1 more. Returns: result
set_device_paramA
Set one parameter on a device or plug-in, found by (fuzzy) name or index.
Give exactly one of:
value: raw value within the parameter's min..max (see get_device_params)
normalized: 0.0-1.0 across the parameter's range
option: a named choice for switch/menu parameters (e.g. "Low Cut") Returns the before/after values including the readable display string.
Inputs: parameter (required), track_index (required), value, option, +4 more. Returns: result
set_device_parameterA
Set a device parameter to a specific value.
Use get_device_parameters first to discover parameter indices and ranges.
Parameters:
track_index: Track that owns the device
device_index: Index into the track's device chain
parameter_index: Index into device.parameters
value: New parameter value (Live parameter units)
user_prompt: The original user prompt that led to this tool call (for telemetry)
Inputs: value (required), track_index (required), device_index (required), parameter_index (required), +1 more. Returns: result
set_locatorB
Create or rename a locator at a beat (reliable: waits for Live's playhead to move first).
Inputs: time_beats (required), name. Returns: result
set_mixerA
Set any of a track's mixer controls in one call; omitted controls are left alone.
volume: 0.0-1.0 fader value (0.85 = 0 dB, 0.0 = -inf)
pan: -1.0 (left) to 1.0 (right)
sends: {"0": 0.5, "1": 0.2} keyed by return index, 0.0-1.0
track_type: "track" (default), "return" or "master" (track_index is ignored for master)
Inputs: arm, pan, mute, solo, +4 more. Returns: result
set_plugin_presetCDestructive
Switch a plug-in to one of its exposed presets, by (fuzzy) name or index.
Inputs: preset (required), track_index (required), track_type, device_name, +1 more. Returns: result
set_routingC
Set routing by (fuzzy) display name, e.g. input_type="Ext. In", input_channel="1".
Inputs: track_index (required), input_type, track_type, output_type, +2 more. Returns: result
set_scene_nameC
Rename a scene (e.g. "Verse", "Chorus").
Inputs: name (required), index (required). Returns: result
set_sidechainB
Route a track (by name, e.g. "Drums") into a Compressor's sidechain input and switch S/C On. Live's API routes sidechains only into Compressor (not Glue Compressor, Gate or Auto Filter).
Inputs: source (required), track_index (required), enable, track_type, +2 more. Returns: result
set_simplerCDestructive
Simpler playback: mode classic / one-shot / slicing; slice_by transient, beat, region, manual; sensitivity 0..1 for transient slicing; warp the sample on/off.
Inputs: track_index (required), mode, warp, voices, +3 more. Returns: result
set_song_scaleA
Set the song key, e.g. root="A", scale_name="Minor"; scale_mode=True highlights it in MIDI clips.
Inputs: root, scale_mode, scale_name. Returns: result
set_tempoB
Set the tempo of the Ableton session.
Parameters:
tempo: The new tempo in BPM
user_prompt: The original user prompt that led to this tool call (for telemetry)
Inputs: tempo (required), user_prompt. Returns: result
set_track_colorB
Set a track's colour by Live palette index (0-69).
Inputs: color_index (required), track_index (required), track_type. Returns: result
set_track_foldB
Fold or unfold a group track.
Inputs: track_index (required), folded. Returns: result
set_track_nameB
Set the name of a track.
Parameters:
track_index: The index of the track to rename
name: The new name for the track
user_prompt: The original user prompt that led to this tool call (for telemetry)
Inputs: name (required), track_index (required), user_prompt. Returns: result
set_transportC
Arrangement record, Session record, overdub, metronome, loop region (beats) and time signature. Arm the tracks first with set_mixer(arm=True).
Inputs: loop, record, overdub, metronome, +5 more. Returns: result
song_starterADestructive
Start a song in a genre in one step: the session template (instruments, chains, returns) plus a first scene with chords, bass, drums and a melody that fit together. Genres: ballad, pop, mizrahi_pop, singer_songwriter, cinematic, house, six_eight. key/tempo override the genre's; change seed for another idea. build_template=False writes the parts into the existing Piano/Bass/Drums tracks instead.
Inputs: key, seed, genre, tempo, +1 more. Returns: result
start_playbackB
Start playing the Ableton session.
Parameters:
user_prompt: The original user prompt that led to this tool call (for telemetry)
Inputs: user_prompt. Returns: result
stop_clipC
Stop playing a clip.
Parameters:
track_index: The index of the track containing the clip
clip_index: The index of the clip slot containing the clip
user_prompt: The original user prompt that led to this tool call (for telemetry)
Inputs: clip_index (required), track_index (required), user_prompt. Returns: result
stop_playbackB
Stop playing the Ableton session.
Parameters:
user_prompt: The original user prompt that led to this tool call (for telemetry)
Inputs: user_prompt. Returns: result
swing_clipADestructive
Swing a Session MIDI clip: delay off-beat 8ths (or 16ths) by amount (0..1 of the grid; 0.33 ≈ triplet feel). For Live's own grooves use set_clip_groove.
Inputs: clip_index (required), track_index (required), grid, amount. Returns: result
switch_to_arrangement_viewB
Switch Ableton's main window to the Arrangement view.
Parameters:
user_prompt: The original user prompt that led to this tool call (for telemetry)
Inputs: user_prompt. Returns: result
transform_clipADestructive
Develop a motif from a Session MIDI clip and write the result to another slot (default: the next empty slot below; the same slot overwrites the source). Operations, all staying in the key: invert, retrograde, augment, diminish (amount = factor, default 2), transpose (amount = scale steps), sequence (repeats copies, each amount steps higher), additive (minimalist build-up), displace (amount = beats). key/scale default to the song's scale.
Inputs: operation (required), clip_index (required), track_index (required), key, +4 more. Returns: result
transpose_clipBDestructive
Transpose a clip: MIDI notes are moved; audio clips get their pitch shifted.
Inputs: semitones (required), clip_index (required), track_index (required), arrangement. Returns: result
undoADestructive
Undo the last N actions in Live (same as Ctrl+Z).
Inputs: steps. Returns: result
write_basslineADestructive
Write a bass part that follows the chords (slash basses respected) into a Session clip. style: whole, root_fifth, pop8, ballad (root + passing note), walking, octaves, arp. low/high: MIDI range (28 = E1, the lowest bass string).
Inputs: clip_index (required), track_index (required), key, low, +8 more. Returns: result
write_chordsBDestructive
Write a chord progression by symbol into a Session clip (created if empty), with smooth voice leading.
chords: ["Am7", "Fmaj7", "C/E", "G7sus4"] — also add9, m7b5, dim7, 6/9, 9, 11, 13...
style: close, open, spread (two-hand piano), drop2, shell, power
pattern: block, half, arp_up, arp_down, arp_updown, broken, pulse8
center: MIDI note the voicing sits around (60 = middle C)
Inputs: chords (required), clip_index (required), track_index (required), style, +7 more. Returns: result
write_drum_patternBDestructive
Write a drum pattern for a Drum Rack (GM layout: kick 36, rim 37, snare 38, hat 42, crash 49). style: ballad, pop, halftime, six_eight, brush_rim, four_floor, fill. notes remaps pads, e.g. {"crash": 48} — get_drum_pads shows what each pad holds.
Inputs: clip_index (required), track_index (required), bars, seed, +5 more. Returns: result
write_euclideanBDestructive
Euclidean rhythms (pulses spread as evenly as possible over steps) into a Session clip. layers: [{"pitch": 36, "pulses": 4, "steps": 16}, {"pitch": 42, "pulses": 7, "steps": 16, "rotation": 2}, {"pitch": 37, "pulses": 3, "steps": 8}] — different step counts make polyrhythms. Classics: 3/8 tresillo, 5/8 cinquillo, 7/16 samba, 5/16 bossa. Pitches are Drum Rack pads (36 kick, 38 snare, 42 hat).
Inputs: layers (required), clip_index (required), track_index (required), bars, +2 more. Returns: result
write_melodyBDestructive
Write a melody from scale degrees (1 = root, 8 = octave up, 0 = rest) and durations in beats. Scales: major, minor, harmonic minor, melodic minor, dorian, phrygian, lydian, mixolydian, locrian, major/minor pentatonic, blues, phrygian dominant.
Inputs: root (required), scale (required), rhythm (required), degrees (required), +5 more. Returns: result
write_named_progressionCDestructive
Write a progression from the library (find_progressions) in any key, e.g. name="andalusian", key="A" -> Am G F E. style: close, open, spread, drop2, shell, rootless, quartal.
Inputs: key (required), name (required), clip_index (required), track_index (required), +5 more. Returns: result
write_progressionCDestructive
Write a progression in roman numerals in any key, e.g. key="A", scale="minor", numerals=["i", "VI", "III", "VII"] or key="C", numerals=["I", "vi7", "IV", "V7", "bVII"].
Inputs: key (required), numerals (required), clip_index (required), track_index (required), +7 more. Returns: result
Tool Schema Changelog
Recent tool additions, removals, and schema changes observed during successful MCP inspections.
Loading tool schema changelog…
TDQS
Scored across 148 tools
There are direct near-duplicate pairs (get_device_params vs get_device_parameters, set_device_param vs set_device_parameter) and many overlapping effect-loading paths (load_instrument_or_effect, insert_device, add_effect, load_plugin, load_browser_uri). Three separate progression-writing tools (write_chords, write_progression, write_named_progression) and multiple mixer/device-setting routes make tool selection error-prone.
Most names are snake_case verb_noun, but noun-only tools (chords, song_starter, session_diff, preset_info, song_starter) and inconsistent verb choices break predictability. The _param vs _parameter duplication and mixed legacy/new naming further muddy the convention.
148 tools is an extreme mismatch for a DAW-control server; even with broad Ableton functionality, the surface is far beyond a reasonable scope. Many niche tools (als_lint, als_diff, inspect_live_api, scan_plugin_catalog, save_plugin_layout) overwhelm the core workflow.
Coverage is very broad: tracks, clips, notes, mixer, devices, plug-ins, arrangement, browser, mixing, export, analysis, and songwriting are all represented. Only minor gaps (e.g., granular per-note editing and some clip lifecycle operations) keep it from perfect completeness.
Maintenance
Related MCP Connectors
Compose, mix, master and export in Logic Pro, REAPER or Ableton Live by talking to an AI engineer.
AI mixing and mastering: analyze your mixes, run DSP autofix, render stems, and master tracks.
Read and edit DAW projects as an agent, save native Logic projects, convert projects between DAWs
Make AI Tracks, loops and sound effects, split stems, build drum kits and master songs.
Related MCP Servers
- AlicenseNot gradedqualityDmaintenanceBridges AI assistants with Ableton Live, enabling real-time control, offline project analysis, version tracking, and rack/preset parsing for music production workflows.MIT
- AlicenseCqualityDmaintenanceConnects Ableton Live to Claude AI, enabling prompt-assisted music production, track creation, and Live session manipulation.100MIT
- AlicenseCqualityBmaintenanceEnables Claude or any AI agent to operate Live 12 Intro end to end: creating tracks and scenes, writing and quantizing MIDI clips, loading audio onto the Arrangement timeline, setting device parameters in real units, drawing clip automation, mixing with real sidechain, and rendering or analyzing audio so the agent can check its own work. It exposes 57 tools over MCP or a local JSON socket, working within the stock Intro edition's 16-track, 16-scene limits.57MIT
- AlicenseAqualityBmaintenanceEnables AI agents and MCP clients to drive Ableton Live 12 end to end, from setting tempo and building instruments through writing MIDI, arranging, mixing, bouncing and mastering. Adds real-time rendering plus loudness and spectral analysis and multi-format encoding with tags and artwork, for turning an empty set into a release-ready track.78MIT