lmms-mcp
This MCP server lets AI assistants create, edit, arrange, mix, and render LMMS song projects; composition works without LMMS running, while rendering requires an LMMS installation.
Manage projects: create, load, save, inspect project info/XML, and set tempo, time signature, master volume, and master pitch.
Manage tracks: add instrument, sample, pattern/Beat-Bassline, automation, and VST tracks; list/get/remove tracks; set volume, pan, mute, and solo.
Add/edit notes and clips: MIDI or note-name notes, batch notes, patterns, sample clips, BB clips; move/delete clips and inspect arrangement.
Mix and process: add/rename/set volume on mixer channels; add/remove/toggle/list effects on tracks or mixer channels.
Automate parameters: song tempo, master volume/pitch, track volume/pan, and mixer volume over time with smooth or linear curves.
Use presets/plugins: list/load ZynAddSubFX presets, set Zyn parameters, list available LMMS plugins, and scan VST directories.
Render/export: render the saved project to WAV, FLAC, OGG, or MP3 via headless LMMS.
Music helpers: convert note names/MIDI keys, convert bars/ticks, and generate scales.
Enables AI assistants to compose, arrange, and mix songs for LMMS by writing LMMS project files (.mmpz/.mmp). Supports tempo and time signature setup, instrument and drum tracks, notes and chords, patterns, arrangement, mixer channels and effects, automation, MIDI import, project editing, rendering to audio, and opening projects in LMMS.
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., "@lmms-mcp Create a new LMMS project, set tempo to 90 BPM, and add a basic drum pattern."
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.
LMMS MCP Server
An MCP server that lets AI assistants compose, arrange and mix songs for LMMS, the free open-source DAW.
The server writes regular LMMS project files (.mmpz / .mmp). You open the result in LMMS, or let the server render it to audio. LMMS does not have to be running while the song is built.
What it can do
Songs from scratch – tempo, time signature, tracks, notes, arrangement, mix
Musical input – note names (
C4,F#3), a compact melody notation ("A1:e A1 r A2 A1:q"), chord symbols ("Am F C G"), step-sequencer strings ("x...x...x...x...")Drums – drum kits from the LMMS factory samples in the Beat/Bassline (pattern) editor, patterns and variations
Instruments – all built-in synths, the sample player, SoundFonts, LMMS presets (
.xpf) and ZynAddSubFX patches (.xiz)Sound shaping – envelopes, LFOs, filters, arpeggiator, chord stacking
Arrangement – repeat or clear whole sections, audio clips on sample tracks, MIDI file import
Mixing – mixer channels, sends, built-in effects with parameters
Automation – tempo, volumes, pan, filter cutoff, any instrument or effect parameter; LFO controllers
Existing projects – opens and edits projects saved by LMMS 1.0 – 1.3 without losing data it does not understand
LMMS 1.2 and 1.3 – writes either file format (see File format)
Safety net – undo, validation ("this track will be silent because…"), failed calls never leave a half-edited project
Related MCP server: LMMS Agent Bridge
Installation
Requires Python 3.10+.
git clone https://github.com/TypeWolf/lmms-mcp.git
cd lmms-mcp
pip install -e .Configuration
Claude Desktop (claude_desktop_config.json)
{
"mcpServers": {
"lmms": {
"command": "python",
"args": ["-m", "lmms_mcp"],
"env": {
"LMMS_PROJECTS_DIR": "C:\\Users\\me\\Documents\\lmms\\projects"
}
}
}
}opencode (opencode.json)
{
"mcp": {
"lmms": {
"type": "local",
"command": ["python", "-m", "lmms_mcp"],
"environment": {
"LMMS_PROJECTS_DIR": "C:\\Users\\me\\Documents\\lmms\\projects"
}
}
}
}Environment variables
All are optional – an LMMS installed in the default location is found automatically.
Variable | Purpose |
| Folder for relative project paths. Default: |
| Path to |
| LMMS data folder containing |
|
|
| LMMS working directory with your own |
| Folder containing the LMMS plugin libraries if they are not found automatically (AppImage or custom install) |
Ask the assistant to call get_environment to see what was detected.
Plugin availability
The server reads the plugin libraries of the installed LMMS when it can find them. Availability has three states: detected, missing from a nonempty inventory, and unknown. An empty or unreadable inventory, as well as a plugin that is only assumed to be statically linked, is not verification.
get_environment reports plugin_inventory_status, instruments_verified_available, effects_verified_available, instruments_unverified and effects_unverified. The verified fields only mean that a matching plugin file was found; they do not prove that LMMS loads the plugin or produces audio. Offline composition remains supported: add_instrument_track and add_effect add a warning when availability cannot be checked or the plugin is missing from a nonempty inventory.
Set LMMS_PLUGIN_DIR when the plugin folder is not in a standard location (for example an AppImage or a custom install). Linux/AppImage discovery remains best-effort.
Example
"Make a 16-bar deep house loop in A minor at 122 BPM and open it in LMMS."
The assistant will typically run something like:
new_project(name="deep-house", tempo=122)
add_drum_kit(kit="electro")
set_beat("Main", {"Kick": "x...x...x...x...", "Clap": "....x.......x...",
"Hihat": "..x...x...x...x.", "Openhat": "..............x."})
arrange_pattern("Main", bar=1, length_bars=16)
add_instrument_track("Bass", instrument="lb302", params={"vcf_cut": 0.4}, new_mixer_channel=True)
add_notes("Bass", sequence="A1:e r A1 r:q A1:e C2 r G1:q.", bar=1)
add_instrument_track("Keys", preset="Rhodes")
add_chords("Keys", "Am7 Fmaj7 Cmaj7 G", bar=1, rhythm="x..x..x...x.....")
copy_section(from_bar=1, length_bars=4, to_bar=5, times=3, tracks=["Bass", "Keys"])
add_effect("reverbsc", track="Keys", wet=0.25)
add_automation("cutoff", track="Keys", points=[[1, 600], [9, 9000]])
save_project()
open_in_lmms()Conventions
Thing | Convention |
Bars | 1-based position on the song timeline (bar 1 = start); fractions allowed |
Note start / length | Beats (quarter notes); start counts from the given bar, 0-based |
Pitch |
|
Velocity, volume | 0–200, 100 = normal (LMMS native) |
Pan | −100 (left) … 100 (right) |
Tracks | Referenced by name or index; mixer channels by number (0 = master) |
Sequence notation (add_notes(sequence=…)): tokens are played one after another.
C4:1 note for one beat · r:0.5 rest · [C4 E4 G4]:2 chord · E4 reuses the previous duration · q./e/s/h/w letter durations · C4:1@120 velocity · | ignored.
Step strings (set_beat): one character per sixteenth – x hit, X accent, o ghost note, . rest.
Tools
Area | Tools |
Project |
|
Tracks |
|
Sound |
|
Notes |
|
Drums |
|
Arrangement |
|
Mixer / FX |
|
Automation |
|
Library |
|
LMMS |
|
Resources: lmms://guide (workflow and sound-design tips), lmms://project (current project).
Prompt: compose_song(style, tempo, key, bars).
File format
LMMS changed its project format between 1.2 and 1.3 (element names, and note numbers shifted by an octave). The server handles both:
1.2– the LMMS 1.2.x layout. LMMS 1.3 upgrades such files automatically when loading, so this is the safe choice and the default when no LMMS installation is detected.1.3– the current layout of LMMS 1.3 (alpha / nightly builds).
When an existing project is opened, it is kept in whatever version it was saved with. Pitches are always given as real MIDI notes; the server converts them.
Effects and instruments that only exist in 1.3 (compressor, lomm, dispersion, slicert, …) are flagged when used in a 1.2 project.
Limitations
LADSPA, LV2 and VST effects of existing projects are preserved but cannot be created (only the built-in effects can).
VST instruments (
vestige) can be referenced by plugin path but not configured.sfxr,organic,monstro,opulenz,nes,freeboyandsidstart with the plugin's default sound;watsyn,vibedstringsandxpressiveneed a preset.Audio clip lengths are detected for wav, ogg and flac; pass
length_beatsfor other formats.Plugin availability is based on plugin files found in the installed LMMS; file presence does not prove loading or audio output.
Development
pip install -e ".[dev]"
pytest
# plugin availability regression tests (no LMMS installation required)
pytest tests/test_plugin_availability.py
# include the tests that load every factory preset and round-trip the LMMS demo songs
LMMS_DATA_DIR=/path/to/lmms/data pytestLayout:
Module | Content |
| MCP tools, resources, prompt |
| The project model – all editing operations on the LMMS XML |
| Differences between LMMS file format versions |
| Instruments, effects, parameters, drum kits |
| Pitches, scales, chords, note / step notation |
| Finding LMMS, samples, presets; rendering |
| Reading / writing |
| Standard MIDI file reader |
License
MIT – see LICENSE.
Available Tools
52 toolsadd_automationA
Create an automation curve that controls a parameter over time.
Automatable targets:
song + tempo: Song BPM (e.g. tempo ramps)
song + master_volume / master_pitch
track + volume / panning (target_index = track index)
mixer + volume (target_index = mixer channel number)
Args: target_type: "song", "track" or "mixer" param: Parameter name (see above) points: Curve points as list of {"bar": float, "value": float}. Example: [{"bar": 0, "value": 120}, {"bar": 8, "value": 140}] target_index: Track index or mixer channel (ignored for "song") name: Automation name (defaults to parameter name) smooth: True = smooth curves (cubic), False = linear steps
Inputs: param (required), points (required), target_type (required), name, +2 more. Returns: result
add_automation_trackC
Add an automation track for parameter automation.
Args: name: Track name
Inputs: name. Returns: result
add_effectA
Add an effect to a track's or mixer channel's effect chain.
Args: target_type: "track" or "mixer" target_index: Track index or mixer channel number (0=Master) effect: Built-in LMMS effect name. Valid: amplifier, bassbooster, bitcrush, compressor, crossovereq, delay, dispersion, dualfilter, dynamicsprocessor, eq, flanger, frequencyshifter, multitapecho, reverbsc, slewdistortion, stereoenhancer, stereomatrix, waveshaper wet: Wet/dry mix 0.0-1.0 (1.0=full effect) enabled: Whether the effect is active position: Chain position to insert at (None=end of chain)
Inputs: effect (required), target_type (required), target_index (required), wet, +2 more. Returns: result
add_instrument_trackA
Add an instrument track to the project.
Args: name: Track name (e.g. "Lead Synth", "Bass") instrument: Plugin name (lowercase). Valid options: tripleoscillator, kicker, audiofileprocessor, organic, malletsstk, freeboy, lb302, monstro, nes, opulenz, patman, sf2player, sfxr, sid, slicert, vibedstrings, watsyn, xpressive, zynaddsubfx, gigplayer, bitinvader mixer_channel: Mixer channel number (0=Master, 1+=custom channels) volume: Track volume (0-200, 100=normal) panning: Track panning (-100 to +100, 0=center)
Inputs: name (required), volume, panning, instrument, +1 more. Returns: result
add_mixer_channelA
Add a new mixer channel.
The channel is automatically routed to Master via an explicit element - LMMS drops the implicit connection when loading, so channels without it would be silent.
Args: name: Channel name (e.g. "Drums", "Bass", "Lead") volume: Channel volume (0.0-2.0, 1.0=0dB)
Inputs: name (required), volume. Returns: result
add_noteA
Add a note to a track's pattern.
Args: track_index: Zero-based track index (must be an instrument or pattern track) key: MIDI note number (0-127). 60=C4 (middle C), 69=A4 pos: Position in ticks RELATIVE TO THE PATTERN START (192 ticks = 1 bar). Notes are placed inside the target pattern; if a note ends beyond the pattern clip length, the clip is extended automatically (notes beyond the clip end would otherwise be silent). length: Note length in ticks (48 = 1/16 note, 96 = 1/8 note, 192 = 1 bar) volume: Note velocity (0-200, 100=normal) panning: Note panning (-100 to +100) pattern_index: Which pattern on the track to edit (default: first). Use place_pattern to create additional patterns at specific song positions first.
Inputs: track_index (required), key, pos, length, +3 more. Returns: result
add_note_by_nameB
Add a note using a note name (e.g. 'C4', 'A#3', 'F#5').
Args: track_index: Zero-based track index note_name: Note name like 'C4', 'A#3', 'F5', 'Bb2' pos: Position in ticks (192 ticks = 1 bar) length: Note length in ticks (48 = 1/16, 96 = 1/8, 192 = 1 bar) volume: Note velocity (0-200)
Inputs: track_index (required), pos, length, volume, +1 more. Returns: result
add_notes_batchA
Add multiple notes to a track at once.
Args: track_index: Zero-based track index notes: List of note objects, each with: key (int or str like "C4"), pos (ticks), length (ticks, default 48), volume (0-200, default 100), panning (-100 to +100, default 0)
Inputs: notes (required), track_index (required). Returns: result
add_pattern_trackA
Add a beat/bassline pattern track for drum sequencing.
A default inner kick instrument is created so notes can be added immediately via add_note. Trigger it in the song editor with place_bb_clip.
Args: name: Pattern track name
Inputs: name. Returns: result
add_sample_trackB
Add a sample track for arranging audio files.
Args: name: Track name mixer_channel: Mixer channel number (0=Master) volume: Track volume (0-200) panning: Track panning (-100 to +100)
Inputs: name (required), volume, panning, mixer_channel. Returns: result
add_vst_trackA
Add a track hosting a VST plugin (.dll file).
Uses LMMS's Vestige host. The VST must be compatible with your LMMS architecture (64-bit LMMS needs 64-bit VSTs).
Args: name: Track name (e.g. "Spire Lead") dll_path: Absolute path to the VST .dll file mixer_channel: Mixer channel number (0=Master) volume: Track volume (0-200) panning: Track panning (-100 to +100)
Inputs: name (required), dll_path (required), volume, panning, +1 more. Returns: result
assign_sample_fileA
Assign an audio file to all clips on a sample track.
Use relative paths for LMMS's built-in samples (e.g. "drums/kick01.ogg") or absolute paths for your own files.
Args: track_index: Index of the sample track file_path: Audio file path (WAV/OGG/MP3/FLAC)
Inputs: file_path (required), track_index (required). Returns: result
bars_to_ticks_converterA
Convert bars to ticks (192 ticks per bar in 4/4 time).
Args: bars: Number of bars
Inputs: bars (required). Returns: result
create_projectA
Create a new empty LMMS project.
Args: bpm: Tempo in beats per minute (10-999) time_sig_numerator: Time signature numerator (e.g. 4 for 4/4) time_sig_denominator: Time signature denominator (e.g. 4 for 4/4) master_volume: Master volume (0-200, 100=normal) master_pitch: Master pitch offset in semitones (-12 to +12)
Inputs: bpm, master_pitch, master_volume, time_sig_numerator, +1 more. Returns: result
delete_clipA
Delete a clip at a given position from a track.
Args: track_index: Track containing the clip pos_bars: Clip start position in bars
Inputs: pos_bars (required), track_index (required). Returns: result
generate_scaleA
Generate a musical scale as MIDI key numbers and note names.
Args: root_note: Root note name (e.g. 'C4', 'A3') scale_type: Scale type: major, minor, dorian, mixolydian, pentatonic_major, pentatonic_minor, blues, chromatic, harmonic_minor, melodic_minor num_octaves: Number of octaves to generate
Inputs: root_note, scale_type, num_octaves. Returns: result
get_arrangementA
Get the full song editor arrangement: all clips sorted by time.
Shows every pattern, BB clip, sample clip and automation curve with their positions and lengths in ticks and bars.
Inputs: no named parameters. Returns: result
get_effect_chainA
List all effects on a track's or mixer channel's effect chain.
Args: target_type: "track" or "mixer" target_index: Track index or mixer channel number
Inputs: target_type (required), target_index (required). Returns: result
get_lmms_infoA
Get info about the installed LMMS application.
Shows the detected LMMS version, installation path and which instrument/effect plugins are actually available. Use this to check whether a plugin is supported before using it.
Inputs: no named parameters. Returns: result
get_mixer_channelsA
Get all mixer channels with their settings.
Inputs: no named parameters. Returns: result
get_project_infoA
Get comprehensive information about the current LMMS project.
Returns tempo, time signature, tracks, mixer channels, and more.
Inputs: no named parameters. Returns: result
get_project_xmlA
Get the raw XML representation of the current project.
Useful for debugging or understanding the exact project structure.
Inputs: no named parameters. Returns: result
get_trackA
Get detailed information about a specific track.
Args: track_index: Zero-based index of the track
Inputs: track_index (required). Returns: result
key_to_note_nameA
Convert a MIDI key number to its note name.
Args: key: MIDI key number (0-127)
Inputs: key (required). Returns: result
list_available_pluginsA
List ALL plugins installed in your LMMS, dynamically detected.
Includes built-in instruments/effects plus any custom plugins the user added to LMMS's plugins folder. Custom plugins can be used directly by name in add_instrument_track / add_effect.
Inputs: no named parameters. Returns: result
list_tracksA
List all tracks in the current project with their index, name, and type.
Inputs: no named parameters. Returns: result
list_zyn_presetsA
List available ZynAddSubFX presets (.xiz files).
Args: category: Optional category filter (e.g. "Bass", "Strings", "Synth", "Pads", "Brass"). Omit to list all categories' presets.
Inputs: category. Returns: result
load_projectA
Load an existing LMMS project file (.mmpz or .mmp).
Args: path: Full path to the LMMS project file
Inputs: path (required). Returns: result
load_zyn_presetA
Load a ZynAddSubFX preset (.xiz) into a zynaddsubfx instrument track.
The track must use the 'zynaddsubfx' instrument. Presets can be referenced by filename (e.g. "Bass 1"), "Category/Name", or full path.
Args: track_index: Index of the target track preset: Preset name, "Category/Name" or absolute path
Inputs: preset (required), track_index (required). Returns: result
move_clipA
Move a clip (pattern/bbtco/sampleclip) to a new time position.
Args: track_index: Track containing the clip old_pos_bars: Current start position in bars new_pos_bars: New start position in bars
Inputs: track_index (required), new_pos_bars (required), old_pos_bars (required). Returns: result
mute_trackA
Mute or unmute a track.
Args: track_index: Zero-based track index muted: True to mute, False to unmute
Inputs: track_index (required), muted. Returns: result
note_name_to_keyA
Convert a note name (e.g. 'C4') to its MIDI key number.
Args: name: Note name like 'C4', 'A#3', 'F#5', 'Bb2'
Inputs: name (required). Returns: result
place_bb_clipA
Place a beat/bassline clip on a pattern track in the song editor.
This triggers the BB pattern (created via add_pattern_track and filled with notes) to play at the given time.
Args: bb_track_index: Index of the pattern (BB) track pos_bars: Start position in bars length_bars: Clip length in bars (default 4 = one BB cycle of 16 steps)
Inputs: bb_track_index (required), pos_bars, length_bars. Returns: result
place_patternA
Place an empty pattern clip on an instrument track (song arrangement).
The pattern starts empty; add notes with add_note (they go into the first pattern) or use this to sketch the arrangement structure first.
Args: track_index: Index of the instrument track pos_bars: Start position in bars (0 = beginning) name: Optional pattern name (defaults to track name) length_bars: Pattern length in bars
Inputs: track_index (required), name, pos_bars, length_bars. Returns: result
place_sample_clipB
Place an audio file clip on a sample track at a given bar.
Args: track_index: Index of the sample track file_path: Audio file path (relative to LMMS samples or absolute) pos_bars: Start position in bars (0 = beginning) length_bars: Clip length in bars
Inputs: file_path (required), track_index (required), pos_bars, length_bars. Returns: result
remove_effectA
Remove an effect from a chain by name or chain position.
Args: target_type: "track" or "mixer" target_index: Track index or mixer channel number effect: Effect name (e.g. "delay") or position (e.g. 0)
Inputs: effect (required), target_type (required), target_index (required). Returns: result
remove_trackA
Remove a track by its index.
Args: track_index: Zero-based index of the track to remove
Inputs: track_index (required). Returns: result
render_projectA
Export the current project to an audio file using the installed LMMS.
This launches LMMS in headless render mode (no GUI). Requires a working LMMS installation. The project is saved first, then rendered.
Args: output_path: Output file path (default: .wav in the same directory) file_format: "wav", "flac", "ogg" or "mp3" samplerate: Sample rate in Hz (44100, 48000, ...) bitrate: Bitrate in kbit/s for lossy formats (ogg/mp3)
Inputs: bitrate, samplerate, file_format, output_path. Returns: result
save_projectA
Save the current LMMS project to a file.
Args: path: File path to save to. If empty, saves to the default projects directory. If only a filename is given (e.g. "song.mmpz"), it's saved in the default directory. compressed: If True, saves as .mmpz (compressed). If False, saves as .mmp (plain XML). By default (None) the format follows the file extension: ".mmp" is always plain XML, everything else compressed. Writing compressed data into a .mmp breaks LMMS and other tools.
Inputs: path, compressed. Returns: result
scan_vst_directoryA
Scan a folder for VST plugin DLLs (.dll files).
Args: directory: Path to scan (e.g. "C:/VSTPlugins") recursive: Include subdirectories
Inputs: directory (required), recursive. Returns: result
set_master_pitchA
Set the master pitch offset.
Args: pitch: Pitch offset in semitones (-12 to +12)
Inputs: pitch (required). Returns: result
set_master_volumeA
Set the master volume.
Args: volume: Master volume (0-200, 100=normal)
Inputs: volume (required). Returns: result
set_mixer_channel_nameB
Rename a mixer channel.
Args: channel_num: Channel number (0=Master) name: New channel name
Inputs: name (required), channel_num (required). Returns: result
set_mixer_channel_volumeA
Set the volume of a mixer channel.
Args: channel_num: Channel number (0=Master) volume: Volume (0.0-2.0, 1.0=0dB)
Inputs: volume (required), channel_num (required). Returns: result
set_tempoA
Set the song tempo (BPM).
Args: bpm: Tempo in beats per minute (10-999)
Inputs: bpm (required). Returns: result
set_time_signatureA
Set the time signature.
Args: numerator: Beats per bar (e.g. 4, 3, 6) denominator: Beat unit (e.g. 4 for quarter notes, 8 for eighth notes)
Inputs: numerator (required), denominator (required). Returns: result
set_track_panningA
Set the panning of a track.
Args: track_index: Zero-based track index panning: Pan value (-100=left, 0=center, 100=right)
Inputs: panning (required), track_index (required). Returns: result
set_track_volumeA
Set the volume of a track.
Args: track_index: Zero-based track index volume: Volume level (0-200, 100=normal)
Inputs: volume (required), track_index (required). Returns: result
set_zyn_paramsA
Set ZynAddSubFX global parameters on a zynaddsubfx track.
All values are 0-127 as in the ZynAddSubFX UI. Only provided parameters are changed.
Args: track_index: Index of the zynaddsubfx track portamento: Portamento amount (0-127) filterfreq: Filter cutoff frequency (0-127) filterq: Filter resonance/Q (0-127) bandwidth: Bandwidth (0-127) fmgain: FM gain (0-127) rescenterfreq: Resonance center frequency (0-127) resbandwidth: Resonance bandwidth (0-127)
Inputs: track_index (required), fmgain, filterq, bandwidth, +4 more. Returns: result
solo_trackA
Solo or unsolo a track.
Args: track_index: Zero-based track index solo: True to solo, False to unsolo
Inputs: track_index (required), solo. Returns: result
ticks_to_bars_converterA
Convert ticks to bars (192 ticks per bar in 4/4 time).
Args: ticks: Number of ticks
Inputs: ticks (required). Returns: result
toggle_effectA
Enable or disable an effect without removing it.
Args: target_type: "track" or "mixer" target_index: Track index or mixer channel number effect: Effect name or chain position enabled: True to enable, False to bypass
Inputs: effect (required), enabled (required), target_type (required), target_index (required). Returns: result
Tool Schema Changelog
Recent tool additions, removals, and schema changes observed during successful MCP inspections.
Loading tool schema changelog…
TDQS
Scored across 52 tools
Most tools map cleanly to a specific resource (tracks, mixer, effects, clips), but add_note and add_note_by_name are near-duplicates with only input-format differences, and get_project_info overlaps with list_tracks and get_mixer_channels. Descriptions help, but the boundaries are not always crisp.
The set overwhelmingly follows a snake_case verb_noun pattern such as add_instrument_track, remove_effect, set_tempo, and list_tracks. Minor deviations exist in the converter utilities (note_name_to_key, bars_to_ticks_converter) and the add_note_by_name / add_notes_batch variants, but the overall convention is predictable.
52 tools is beyond the 50+ extreme threshold in the rubric, making the surface unwieldy for agent selection. Several tools such as add_note_by_name and the standalone converters could be consolidated into more general tools.
Creation and arrangement workflows are broadly covered, but note editing is incomplete: there is no way to list, update, or delete notes once added, which creates dead ends for pattern editing. Mixer channels can be added but not removed, and effect parameters cannot be adjusted beyond enabling/disabling.
Maintenance
Related MCP Connectors
Read and edit DAW projects as an agent, save native Logic projects, convert projects between DAWs
Create and edit images and video, generate music, and retrieve existing audio projects.
Productivity toolbox for Personal Agents: create PPTs, covers, websites, music and process files.
Autonomous music production for AI agents with MIDI generation, QC and provenance.
Related MCP Servers
- AlicenseAqualityCmaintenanceEnables an AI agent to compose chiptune music by directly reading and writing Lovely Composer's .jsonl project files, including melodies, bass, drums, chords, tempo, and loop points.121MIT
- AlicenseNot gradedqualityCmaintenanceEnables MCP-capable agents to control a running LMMS session to compose, arrange, mix, and export music, including creating tracks, editing notes, loading presets and samples, automating parameters, and rendering audio or MIDI.GPL 2.0
- AlicenseNot gradedqualityBmaintenanceEnables an AI agent to compose and mix directly inside FL Studio: writing and clearing notes, setting tempo, levels and plugin parameters, inspecting what FL currently holds, checking which channels sound, and rendering the arrangement to WAV.3MIT
- FlicenseNot gradedqualityBmaintenanceEnables agents to read, patch, and render DAW projects (REAPER, FL Studio, .rpp/.dawproject) and to measure WAV audio metrics such as peak, RMS, LUFS, and stereo width. It closes the loop by applying patches, rendering before/after audio, and scoring the results against an intent so music-production models can be trained and verified.-