ableton-live-mcp-control

Compare original and translation side by side

🇺🇸

Original

English

🇨🇳

Translation

Chinese

Ableton Live MCP Control

Ableton Live MCP 控制

Skill by ara.so — MCP Skills collection.

This MCP server enables AI agents to control Ableton Live through arbitrary Python execution in Ableton's Live Object Model. It provides both high-level tools for common tasks and low-level access to the entire Ableton API. The server allows creating MIDI clips, inserting audio files, adding tracks/devices, analyzing audio signals, generating spectrograms, and automating mixing/mastering workflows.

由ara.so开发的Skill——MCP Skills合集。

这个MCP服务器允许AI Agent通过在Ableton的Live Object Model中执行任意Python代码来控制Ableton Live。它既提供了用于常见任务的高级工具，也提供了对整个Ableton API的底层访问权限。该服务器支持创建MIDI片段、插入音频文件、添加轨道/设备、分析音频信号、生成频谱图以及自动化混音/母带制作工作流。

Installation

安装

The MCP server requires Ableton Live (tested with Suite 12.3.8) on macOS or Windows.

Automated Setup (Recommended):

Set up the https://github.com/bschoepke/ableton-live-mcp MCP server for me

Manual Setup:

Clone the repository:

bash

git clone https://github.com/bschoepke/ableton-live-mcp.git
cd ableton-live-mcp

Install dependencies:

bash

pip install -r requirements.txt

Install the Ableton Live Remote Script:

Copy the

RemoteScripts/MCP

folder to Ableton's Remote Scripts directory:

macOS:

~/Music/Ableton/User Library/Remote Scripts/

Windows:

%USERPROFILE%\Documents\Ableton\User Library\Remote Scripts\

Configure Ableton Live:
- Open Ableton Live
- Go to Preferences → Link/Tempo/MIDI
- Add "MCP" as a Control Surface
Add to your MCP settings file (e.g.,
```
claude_desktop_config.json
```
):

json

{
  "mcpServers": {
    "ableton-live": {
      "command": "python",
      "args": ["-m", "ableton_live_mcp"],
      "cwd": "/path/to/ableton-live-mcp"
    }
  }
}

Important: Back up your Live Set before using this MCP, as it can edit your set directly.

MCP服务器需要在macOS或Windows系统上运行Ableton Live（已在Suite 12.3.8版本测试）。

自动安装（推荐）：

Set up the https://github.com/bschoepke/ableton-live-mcp MCP server for me

手动安装：

克隆仓库：

bash

git clone https://github.com/bschoepke/ableton-live-mcp.git
cd ableton-live-mcp

安装依赖：

bash

pip install -r requirements.txt

安装Ableton Live远程脚本：

将

RemoteScripts/MCP

文件夹复制到Ableton的远程脚本目录：

macOS：

~/Music/Ableton/User Library/Remote Scripts/

Windows：

%USERPROFILE%\Documents\Ableton\User Library\Remote Scripts\

配置Ableton Live：
- 打开Ableton Live
- 进入偏好设置 → 链接/速度/MIDI
- 添加“MCP”作为控制表面
添加到你的MCP配置文件（例如
```
claude_desktop_config.json
```
）：

json

{
  "mcpServers": {
    "ableton-live": {
      "command": "python",
      "args": ["-m", "ableton_live_mcp"],
      "cwd": "/path/to/ableton-live-mcp"
    }
  }
}

重要提示： 使用此MCP前请备份你的Live工程，因为它可以直接编辑你的工程文件。

Key MCP Tools

核心MCP工具

evaluate_python

Execute arbitrary Python code in Ableton Live's environment.

python

undefined

在Ableton Live环境中执行任意Python代码。

python

undefined

Example: Get the current tempo

示例：获取当前速度

result = evaluate_python("Live.Song.Song.tempo")

Example: Create a new MIDI track

示例：创建新的MIDI轨道

code = """ song = Live.Song.Song track = song.create_midi_track(-1) track.name = 'My New Track' """ evaluate_python(code)

undefined

code = """ song = Live.Song.Song track = song.create_midi_track(-1) track.name = 'My New Track' """ evaluate_python(code)

undefined

create_midi_clip

Create a MIDI clip with notes on a specified track.

python

undefined

在指定轨道上创建带有音符的MIDI片段。

python

undefined

Create a simple C major chord progression

创建简单的C大调和弦进行

create_midi_clip( track_index=0, clip_slot=0, length_bars=4, notes=[ {"pitch": 60, "start": 0.0, "duration": 1.0, "velocity": 100}, # C {"pitch": 64, "start": 0.0, "duration": 1.0, "velocity": 100}, # E {"pitch": 67, "start": 0.0, "duration": 1.0, "velocity": 100}, # G ] )

undefined

undefined

insert_audio_file

Insert an audio file into a track.

python

undefined

将音频文件插入到轨道中。

python

undefined

Insert a vocal sample

插入人声采样

insert_audio_file( track_index=1, clip_slot=0, file_path="/path/to/vocal.wav" )

undefined

insert_audio_file( track_index=1, clip_slot=0, file_path="/path/to/vocal.wav" )

undefined

add_track

Add a new audio or MIDI track with optional devices.

python

undefined

添加带有可选设备的新音频或MIDI轨道。

python

undefined

Add a MIDI track with a synth

添加带有合成器的MIDI轨道

add_track( track_type="midi", name="Bass Synth", devices=["Operator", "Reverb"] )

undefined

add_track( track_type="midi", name="Bass Synth", devices=["Operator", "Reverb"] )

undefined

get_live_set_info

Get comprehensive information about the current Live Set.

python

undefined

获取当前Live工程的全面信息。

python

undefined

Get full Live Set details

获取完整的Live工程详情

info = get_live_set_info()

Returns: tracks, scenes, tempo, time signature, devices, etc.

返回内容：轨道、场景、速度、拍号、设备等

undefined

undefined

capture_audio

Capture audio from an "Agent Audio Tap" Max for Live device.

python

undefined

从“Agent Audio Tap”Max for Live设备捕获音频。

python

undefined

Capture 10 seconds of audio from track 0

从轨道0捕获10秒音频

audio_data = capture_audio( track_index=0, device_index=0, duration_seconds=10.0 )

Returns: base64-encoded WAV data

返回：base64编码的WAV数据

undefined

undefined

get_device_parameters

Get all parameters for a device on a track.

python

undefined

获取轨道上某个设备的所有参数。

python

undefined

Get parameters for the first device on track 0

获取轨道0上第一个设备的参数

params = get_device_parameters( track_index=0, device_index=0 )

undefined

params = get_device_parameters( track_index=0, device_index=0 )

undefined

set_device_parameter

Set a device parameter value.

python

undefined

设置设备参数值。

python

undefined

Set the cutoff frequency of a filter

设置滤波器的截止频率

set_device_parameter( track_index=0, device_index=0, parameter_name="Filter Freq", value=1200.0 )

undefined

set_device_parameter( track_index=0, device_index=0, parameter_name="Filter Freq", value=1200.0 )

undefined

Common Patterns

常见使用模式

Creating a Complete Track with Instruments

创建包含乐器的完整轨道

python

undefined

python

undefined

Create a bass track with Operator synth

创建带有Operator合成器的贝斯轨道

code = """ song = Live.Song.Song track = song.create_midi_track(-1) track.name = 'Bass'

Add Operator synth

添加Operator合成器

operator = track.devices[0] # Default instrument operator.name = 'Bass Synth'

operator = track.devices[0] # 默认乐器 operator.name = 'Bass Synth'

Create a bass pattern

创建贝斯节奏型

clip = track.clip_slots[0].create_clip(4.0) clip.name = 'Bass Pattern'

Add notes

添加音符

for i in range(16): if i % 4 == 0: clip.set_notes(((36, i * 0.25, 0.25, 100, False),)) # C1 on downbeats """ evaluate_python(code)

undefined

for i in range(16): if i % 4 == 0: clip.set_notes(((36, i * 0.25, 0.25, 100, False),)) # 重拍位置的C1 """ evaluate_python(code)

undefined

Analyzing Audio with Spectrograms

用频谱图分析音频

python

undefined

python

undefined

First, add Agent Audio Tap device to a track in Ableton

首先，在Ableton的轨道中添加Agent Audio Tap设备

Then capture and analyze audio

然后捕获并分析音频

import base64 import numpy as np import matplotlib.pyplot as plt from scipy import signal

Capture audio

捕获音频

audio_b64 = capture_audio(track_index=0, device_index=0, duration_seconds=10.0)

Decode audio

解码音频

audio_bytes = base64.b64decode(audio_b64)

Generate spectrogram

生成频谱图

code = """ import wave import numpy as np from scipy import signal

with wave.open('temp.wav', 'rb') as wf: audio_data = np.frombuffer(wf.readframes(wf.getnframes()), dtype=np.int16) sample_rate = wf.getframerate()

f, t, Sxx = signal.spectrogram(audio_data, sample_rate)

code = """ import wave import numpy as np from scipy import signal

with wave.open('temp.wav', 'rb') as wf: audio_data = np.frombuffer(wf.readframes(wf.getnframes()), dtype=np.int16) sample_rate = wf.getframerate()

f, t, Sxx = signal.spectrogram(audio_data, sample_rate)

Save or display spectrogram

保存或显示频谱图

"""

undefined

"""

undefined

Setting Up a Mastering Chain

设置母带处理链

python

undefined

python

undefined

Create master processing chain

创建母带处理链

code = """ song = Live.Song.Song master = song.master_track

Add EQ Eight

添加EQ Eight

eq = master.devices.append(song.create_device('EQ Eight'))

Add Compressor

添加压缩器

comp = master.devices.append(song.create_device('Compressor')) comp.parameters[0].value = -12.0 # Threshold

comp = master.devices.append(song.create_device('Compressor')) comp.parameters[0].value = -12.0 # 阈值

Add Limiter

添加限制器

limiter = master.devices.append(song.create_device('Limiter')) limiter.parameters[1].value = -0.3 # Ceiling """ evaluate_python(code)

undefined

limiter = master.devices.append(song.create_device('Limiter')) limiter.parameters[1].value = -0.3 # 上限 """ evaluate_python(code)

undefined

Creating MIDI Clips Programmatically

程序化创建MIDI片段

python

undefined

python

undefined

Generate a chord progression

生成和弦进行

code = """ import Live

song = Live.Song.Song track = song.tracks[0] clip = track.clip_slots[0].create_clip(8.0)

code = """ import Live

song = Live.Song.Song track = song.tracks[0] clip = track.clip_slots[0].create_clip(8.0)

Chord progression: C - Am - F - G

和弦进行：C - Am - F - G

chords = [ [(60, 64, 67), 0.0], # C major [(57, 60, 64), 2.0], # A minor [(53, 57, 60), 4.0], # F major [(55, 59, 62), 6.0], # G major ]

notes = [] for chord, start_beat in chords: for pitch in chord: notes.append((pitch, start_beat, 2.0, 100, False))

clip.set_notes(tuple(notes)) clip.name = 'Chord Progression' """ evaluate_python(code)

undefined

chords = [ [(60, 64, 67), 0.0], # C大调 [(57, 60, 64), 2.0], # A小调 [(53, 57, 60), 4.0], # F大调 [(55, 59, 62), 6.0], # G大调 ]

notes = [] for chord, start_beat in chords: for pitch in chord: notes.append((pitch, start_beat, 2.0, 100, False))

clip.set_notes(tuple(notes)) clip.name = 'Chord Progression' """ evaluate_python(code)

undefined

Automating Mixing Parameters

自动化混音参数

python

undefined

python

undefined

Create automation for volume fadeout

创建音量淡出自动化

code = """ song = Live.Song.Song track = song.tracks[0] clip = track.clip_slots[0].clip

Create volume automation

创建音量自动化

envelope = clip.automation_envelope(track.mixer_device.volume) envelope.insert_step(0.0, 0, 1.0) # Start at full volume envelope.insert_step(7.0, 0, 0.0) # Fade to silence at bar 7 """ evaluate_python(code)

undefined

envelope = clip.automation_envelope(track.mixer_device.volume) envelope.insert_step(0.0, 0, 1.0) # 开始时音量最大 envelope.insert_step(7.0, 0, 0.0) # 第7小节时淡出至静音 """ evaluate_python(code)

undefined

Working with Audio Effects

音频效果使用

python

undefined

python

undefined

Add and configure a reverb chain

添加并配置混响链

code = """ song = Live.Song.Song track = song.tracks[1]

Add Reverb

添加混响

reverb = track.devices.append(song.create_device('Reverb')) reverb.parameters[0].value = 3.5 # Decay time reverb.parameters[1].value = 0.3 # Dry/Wet

reverb = track.devices.append(song.create_device('Reverb')) reverb.parameters[0].value = 3.5 # 衰减时间 reverb.parameters[1].value = 0.3 # 干湿比

Add EQ after reverb

在混响后添加EQ

eq = track.devices.append(song.create_device('EQ Eight'))

High-pass filter to clean up reverb

高通滤波器清理混响低频

eq.parameters[1].value = 300.0 # Frequency """ evaluate_python(code)

undefined

eq.parameters[1].value = 300.0 # 频率 """ evaluate_python(code)

undefined

Sidechain Compression Setup

侧链压缩设置

python

undefined

python

undefined

Set up sidechain compression (kick to bass)

设置侧链压缩底鼓控制贝斯

code = """ song = Live.Song.Song bass_track = song.tracks[1] kick_track = song.tracks[0]

Add Compressor to bass track

为贝斯轨道添加压缩器

comp = bass_track.devices.append(song.create_device('Compressor'))

Enable sidechain

启用侧链

comp.parameters[9].value = 1.0 # Sidechain enabled

comp.parameters[9].value = 1.0 # 侧链启用

Set sidechain source to kick track

设置侧链源为底鼓轨道

available_routings = comp.available_input_routing_types for routing in available_routings: if 'Audio From' in routing.display_name and kick_track.name in routing.display_name: comp.input_routing_type = routing break

Configure compression

配置压缩参数

comp.parameters[0].value = -18.0 # Threshold comp.parameters[1].value = 4.0 # Ratio comp.parameters[2].value = 0.1 # Attack comp.parameters[3].value = 0.2 # Release """ evaluate_python(code)

undefined

comp.parameters[0].value = -18.0 # 阈值 comp.parameters[1].value = 4.0 # 比率 comp.parameters[2].value = 0.1 # 启动时间 comp.parameters[3].value = 0.2 # 释放时间 """ evaluate_python(code)

undefined

Working with Third-Party Plugins

第三方插件使用

python

undefined

python

undefined

Add VST/AU instruments

添加VST/AU乐器

code = """ song = Live.Song.Song track = song.tracks[0]

Add Serum (example)

添加Serum示例

serum = track.devices.append(song.create_device('Serum'))

Add Keyscape for piano

添加Keyscape钢琴

keyscape = track.devices.append(song.create_device('Keyscape')) """ evaluate_python(code)

undefined

keyscape = track.devices.append(song.create_device('Keyscape')) """ evaluate_python(code)

undefined

Querying Live Set Information

查询Live工程信息

python

undefined

python

undefined

Get detailed information about all tracks

获取所有轨道的详细信息

info = get_live_set_info()

The response includes:

返回内容包括：

- tempo

- 速度

- time_signature

- 拍号

- tracks (name, type, devices, clip_slots)

- 轨道名称、类型、设备、片段槽

- scenes

- 场景

- master_track info

- 主轨道信息

- return_tracks

- 返回轨道

Example: Find all tracks with a specific device

示例：查找所有带有特定设备的轨道

code = """ song = Live.Song.Song tracks_with_reverb = []

for i, track in enumerate(song.tracks): for device in track.devices: if 'Reverb' in device.name: tracks_with_reverb.append((i, track.name)) break

result = tracks_with_reverb """ result = evaluate_python(code)

undefined

code = """ song = Live.Song.Song tracks_with_reverb = []

for i, track in enumerate(song.tracks): for device in track.devices: if 'Reverb' in device.name: tracks_with_reverb.append((i, track.name)) break

result = tracks_with_reverb """ result = evaluate_python(code)

undefined

Audio Signal Analysis

音频信号分析

python

undefined

python

undefined

Analyze frequency content of a track

分析轨道的频率内容

code = """ import numpy as np from scipy import signal as sp_signal

Assuming audio is captured via Agent Audio Tap

假设音频是通过Agent Audio Tap设备捕获的

(audio_data would be from capture_audio tool)

audio_data来自capture_audio工具

def analyze_frequency_content(audio_data, sample_rate=44100): # Compute FFT fft = np.fft.rfft(audio_data) freqs = np.fft.rfftfreq(len(audio_data), 1/sample_rate) magnitudes = np.abs(fft)

# Find dominant frequencies
peaks, _ = sp_signal.find_peaks(magnitudes, height=np.max(magnitudes)*0.1)
dominant_freqs = freqs[peaks]

return {
    'dominant_frequencies': dominant_freqs.tolist(),
    'frequency_range': [freqs[0], freqs[-1]],
    'peak_magnitude': float(np.max(magnitudes))
}

def analyze_frequency_content(audio_data, sample_rate=44100): # 计算FFT fft = np.fft.rfft(audio_data) freqs = np.fft.rfftfreq(len(audio_data), 1/sample_rate) magnitudes = np.abs(fft)

# 查找主导频率
peaks, _ = sp_signal.find_peaks(magnitudes, height=np.max(magnitudes)*0.1)
dominant_freqs = freqs[peaks]

return {
    'dominant_frequencies': dominant_freqs.tolist(),
    'frequency_range': [freqs[0], freqs[-1]],
    'peak_magnitude': float(np.max(magnitudes))
}

Use this function after capturing audio

捕获音频后使用此函数

"""

undefined

"""

undefined

Configuration

配置

The MCP server connects to Ableton via OSC (Open Sound Control) on localhost.

Default settings:

OSC receive port: 11000
OSC send port: 11001

To change ports, modify the Remote Script configuration in Ableton or update the MCP server settings.

Environment Variables:

No specific environment variables required
Ensure Python path includes the MCP module directory

MCP服务器通过本地主机上的OSC（Open Sound Control）连接到Ableton。

默认设置：

OSC接收端口：11000
OSC发送端口：11001

如需修改端口，请在Ableton中修改远程脚本配置或更新MCP服务器设置。

环境变量：

无需特定环境变量
确保Python路径包含MCP模块目录

Troubleshooting

故障排除

MCP Not Connecting to Ableton

MCP无法连接到Ableton

Verify the MCP Remote Script is installed in Ableton's Remote Scripts folder
Check that "MCP" is selected as a Control Surface in Ableton Preferences
Restart Ableton Live after installing the Remote Script
Check firewall settings allow local OSC communication

确认MCP远程脚本已安装在Ableton的远程脚本文件夹中
检查Ableton偏好设置中已选择“MCP”作为控制表面
安装远程脚本后重启Ableton Live
检查防火墙设置允许本地OSC通信

Evaluation Errors

执行错误

python

undefined

python

undefined

Always wrap risky code in try/except when using evaluate_python

使用evaluate_python时始终将风险代码包裹在try/except中

code = """ try: song = Live.Song.Song # Your code here except Exception as e: result = f'Error: {str(e)}' """

undefined

code = """ try: song = Live.Song.Song # 你的代码 except Exception as e: result = f'Error: {str(e)}' """

undefined

Audio Capture Issues

音频捕获问题

Ensure "Agent Audio Tap" Max for Live device is added to the track
Verify the device is enabled and receiving audio
Check that the track is not muted
Confirm audio is actually playing during capture

确认轨道中已添加“Agent Audio Tap”Max for Live设备
验证设备已启用并正在接收音频
检查轨道未被静音
确认捕获期间音频正在播放

Device Not Found

设备未找到

python

undefined

python

undefined

List all devices on a track first

先列出轨道上的所有设备

code = """ song = Live.Song.Song track = song.tracks[0] device_names = [d.name for d in track.devices] result = device_names """ devices = evaluate_python(code)

undefined

code = """ song = Live.Song.Song track = song.tracks[0] device_names = [d.name for d in track.devices] result = device_names """ devices = evaluate_python(code)

undefined

Parameter Changes Not Taking Effect

参数修改未生效

Ensure parameter names match exactly (case-sensitive)
Some parameters may have limited ranges - check min/max values
Verify the device is enabled
Use
```
get_device_parameters()
```
to see available parameters

确保参数名称完全匹配区分大小写
部分参数可能有范围限制，请检查最小值/最大值
验证设备已启用
使用
```
get_device_parameters()
```
查看可用参数

Live Set Corruption Prevention

防止Live工程损坏

Always back up your Live Set before extensive MCP operations
Test code on a duplicate/test project first
Use version control for your Live Sets
Save frequently during AI-assisted production sessions

在进行大量MCP操作前始终备份你的Live工程
先在副本/测试项目上测试代码
为你的Live工程使用版本控制
在AI辅助制作会话期间频繁保存

Advanced Usage

高级用法

Custom Python Libraries

自定义Python库

You can import and use additional Python libraries in

evaluate_python

python

code = """
import numpy as np
import math

你可以在

evaluate_python

中导入并使用额外的Python库：

python

code = """
import numpy as np
import math

Generate MIDI notes based on harmonic series

基于泛音列生成MIDI音符

fundamental = 220 # A3 harmonics = [fundamental * i for i in range(1, 9)]

Convert to MIDI note numbers

转换为MIDI音符编号

midi_notes = [int(round(69 + 12 * math.log2(f / 440))) for f in harmonics] result = midi_notes """ notes = evaluate_python(code)

undefined

midi_notes = [int(round(69 + 12 * math.log2(f / 440))) for f in harmonics] result = midi_notes """ notes = evaluate_python(code)

undefined

Batch Processing Multiple Tracks

批量处理多个轨道

python

undefined

python

undefined

Apply same effect chain to multiple tracks

为多个轨道应用相同的效果链

code = """ song = Live.Song.Song effect_chain = ['EQ Eight', 'Compressor', 'Reverb']

for i in range(4): # First 4 tracks track = song.tracks[i] for effect_name in effect_chain: track.devices.append(song.create_device(effect_name)) """ evaluate_python(code)

undefined

code = """ song = Live.Song.Song effect_chain = ['EQ Eight', 'Compressor', 'Reverb']

for i in range(4): # 前4个轨道 track = song.tracks[i] for effect_name in effect_chain: track.devices.append(song.create_device(effect_name)) """ evaluate_python(code)

undefined