Gridshift includes a built-in MCP server that allows AI assistants to read project state and control the DAW — including mixing, plugin management, and MIDI composition.
Installing for Claude Desktop
Go to Help → Install Claude Extension in Gridshift’s menu bar. This installs the MCP server as an extension in Claude Desktop.
Adding to Other Clients
The MCP server uses a bridge binary that communicates with Gridshift via Unix socket. Add the following to your client’s MCP configuration:
{
"mcpServers": {
"gridshift": {
"command": "/Applications/Gridshift.app/Contents/Helpers/GridshiftBridge"
}
}
}
Gridshift must be running for the connection to work.
Session
| Tool | Description |
|---|
get_session_overview | Get project info, transport state, tracks, and mixer levels in one call. Use the include parameter to request specific sections. |
Transport
| Tool | Description |
|---|
transport_control | Control playback, recording, loop, tempo — all via a single action parameter (play, stop, toggle_playback, toggle_loop, toggle_recording, go_to_beginning, set_tempo). |
Tracks
| Tool | Description |
|---|
track_add | Add a sample, instrument, or return track. |
track_configure | Set volume, pan, mute, solo, and return output routing for one or more tracks in a single call. Accepts a batch operations array. |
track_rename | Rename a track. |
group_create | Create a group track containing existing tracks as children. |
track_move_to_group | Move a track into a group, or back to the top level. |
Sends & Returns
| Tool | Description |
|---|
send_configure | Create, adjust, or remove sends from tracks to return tracks (level + pre/post tap). Accepts a batch operations array. Create a return track first with track_add type return. |
Plugins
| Tool | Description |
|---|
available_plugins | Search installed AU plugins by name or manufacturer. |
plugins_list | List all loaded plugins across tracks with bypass state and parameter count. |
plugin_get_parameters | Get plugin parameters with type metadata, value labels, and pagination. |
plugin_configure | Set parameter values, toggle bypass, and/or configure sidechain routing — all in a single call. |
plugin_add_remove | Add or remove plugins on tracks. Supports batch operations across multiple tracks. |
Clips & MIDI
| Tool | Description |
|---|
clip_delete | Delete clips by ID or clear all clips on a track. Works with any clip type. |
midi_clip_create | Create a MIDI clip with optional initial notes. Returns the clip ID for further editing. |
midi_clip_read | Read all notes from a MIDI clip (pitch, position, duration, velocity). |
midi_clip_edit | Add, delete, or modify notes in a MIDI clip. Supports batch operations. |
Automation
| Tool | Description |
|---|
automation_create | Automate a parameter (mixer volume/pan, a plugin parameter, or a send amount) over a time range. Creates the lane and clip, and optionally seeds breakpoints or an LFO. absolute sets the value directly; relative stacks a modulation layer on top — so you can drive one parameter from several automations. Returns the lane index and clip ID. |
automation_clip_read | Read an automation clip: its targets, blend mode and range, and breakpoints or LFO shape. |
automation_clip_edit | Add, delete, modify, or re-curve breakpoints in an automation clip. Supports batch operations. |
automation_lfo | Configure LFO generate mode on a clip: set the shape, bake it to editable points, or clear it. |
automation_target_configure | Edit a lane’s parameter bindings: add or remove targets (one curve → many parameters), or change a target’s blend mode or range. |
Samples
| Tool | Description |
|---|
sample_search | Search the indexed sample library by filename, folder, and category tags. |
sample_place | Place a sample as a clip, creating a sample track if needed. |
sample_replace_on_track | Swap the audio source on a sample track, preserving clip positions. |
Project
| Tool | Description |
|---|
project_artwork | Get, set, or remove the project artwork (cover image). Supports PNG and JPEG. |
Design Principles
Gridshift’s MCP tools follow best practices for LLM integration:
- Workflow-oriented: Tools map to user goals, not internal APIs.
- Compound operations: Batch changes in single round-trips (e.g. configure multiple tracks at once).
- Rich metadata: Plugin parameters include type info and human-readable labels so the AI can set values correctly without trial-and-error.
- Musical context: Project key and scale are included in the session overview so the AI writes in-key MIDI.
Last modified on June 5, 2026