SV1 Agent Bridge
A local MCP server that enables AI agents to control Synthesizer V Studio 1.x via its Lua scripting API.
README
SV1 Agent Bridge
A local MCP server that lets Codex (or any MCP client) inspect and control Synthesizer V Studio 1.x through its public Lua scripting API.
Inspired by zhoupengjie/synthv-agent-bridge (which targets Synthesizer V Studio 2 Pro). This version targets Studio 1.x (e.g. 1.11.2) and reuses the same proven pattern: a persistent Lua script that polls JSON request files written by a stdio MCP server.
Codex (stdio MCP) --JSON-RPC--> Node MCP server --file IPC--> SV1AgentBridge.lua (in SV1) --SV API--> project
No network ports, no .svp parsing, no AI calls. All project mutations run
inside Synthesizer V through its scripting object model and create undo records.
Layout
synthv/SV1AgentBridge.luapersistent Lua bridge (install into SV1 Scripts folder)synthv/StopSV1AgentBridge.luastops a running bridgesrc/server.mjszero-dependency stdio MCP server (Node >= 18)ipc/request/response/status files (auto-created)
Install
- Copy
synthv/SV1AgentBridge.luaandsynthv/StopSV1AgentBridge.luainto your Synthesizer V Studio 1 Scripts directory (e.g.Documents\Synthesizer V Studio\Scriptsor the app's Scripts folder). - Keep
src/server.mjshere. Node >= 18 is required (no npm install needed).
Run
- In Synthesizer V Studio 1, open a project, then run SV1 Agent Bridge from the Scripts menu. A message box confirms it started and shows the IPC dir.
- Register the MCP server with Codex (see below), then ask Codex to use it.
MCP config
Codex CLI - add to ~/.codex/config.toml (or the app's MCP settings):
[mcp_servers.sv1-agent-bridge]
command = "node"
args = ["E:/Project/sv1-agent-bridge/src/server.mjs"]
Or via CLI: codex mcp add sv1-agent-bridge node E:/Project/sv1-agent-bridge/src/server.mjs
Claude Code - one command:
claude mcp add sv1-agent-bridge -- node E:/Project/sv1-agent-bridge/src/server.mjs
Cursor, Windsurf, Cline, and generic stdio - see docs/MCP-CONFIG.md.
Optional: set a custom IPC directory with the env var SV1_BRIDGE_IPC on both
sides (the Lua script reads the same env var).
Tools (29 total)
Project / structure
sv1_ping, sv1_read_project, sv1_create_track, sv1_create_note_group,
sv1_add_group_reference, sv1_set_tempo, sv1_set_time_signature
Notes
sv1_read_notes, sv1_get_note_details, sv1_set_lyrics, sv1_edit_note,
sv1_edit_notes, sv1_add_note, sv1_add_notes, sv1_delete_note,
sv1_transpose_notes, sv1_set_musical_type, sv1_refresh_phonemes
Editor / selection
sv1_set_current_track, sv1_set_current_group, sv1_get_selection,
sv1_select_notes, sv1_clear_selection
Playback
sv1_play, sv1_stop, sv1_pause, sv1_seek, sv1_set_loop, sv1_get_state
All track/group/note indices are 1-based. Time values are in blicks
(SV's internal unit; SV.QUARTER = 705600000 blicks per quarter note). Tools
that return notes also include seconds via the project time axis.
sv1_add_notes accepts an optional unit of quarters or blicks.
edit_note/edit_notes/add_note/add_notes accept an optional musicalType
("sing" or "rap").
Known limitation: voice assignment
SV1 1.11.2 scripting cannot assign or copy a voice database to a group
reference: getVoice() returns an empty table, setVoice() crashes the host,
and there is no voice-database enumeration API. New groups created via
sv1_create_note_group + sv1_add_group_reference have no voice and will not
render until you assign one manually in the SV1 UI (select the group, choose a
voice database). Writing notes to a group that already has a voice (assigned in
the UI) works normally.
Documentation
- AGENTS.md - usage manual for any agent driving SV1 through this bridge
- docs/DEVELOPING.md - architecture, IPC protocol, how to add tools, SV1 API gotchas
- docs/MCP-CONFIG.md - MCP server config for Claude Code, Cursor, Windsurf, Cline, etc.
Status
Experimental. Test on a copy of important projects. Write tools create one Synthesizer V undo record each, so you can undo any change from the editor.
How the SV1 vs SV2 gap was handled
Studio 1.x lacks Studio 2 features (Vocal Modes, Smart Pitch, automation
curves, native sidebar). The core scripting model
(SV:getProject/getPlayback/getMainEditor, setTimeout polling, io file
access, Project/Track/GroupReference/NoteGroup/Note) is present and nearly
identical, so the bridge exposes note lyrics/pitch/timing/phonemes + playback.
Recommended Servers
playwright-mcp
A Model Context Protocol server that enables LLMs to interact with web pages through structured accessibility snapshots without requiring vision models or screenshots.
Magic Component Platform (MCP)
An AI-powered tool that generates modern UI components from natural language descriptions, integrating with popular IDEs to streamline UI development workflow.
Audiense Insights MCP Server
Enables interaction with Audiense Insights accounts via the Model Context Protocol, facilitating the extraction and analysis of marketing insights and audience data including demographics, behavior, and influencer engagement.
VeyraX MCP
Single MCP tool to connect all your favorite tools: Gmail, Calendar and 40 more.
graphlit-mcp-server
The Model Context Protocol (MCP) Server enables integration between MCP clients and the Graphlit service. Ingest anything from Slack to Gmail to podcast feeds, in addition to web crawling, into a Graphlit project - and then retrieve relevant contents from the MCP client.
Kagi MCP Server
An MCP server that integrates Kagi search capabilities with Claude AI, enabling Claude to perform real-time web searches when answering questions that require up-to-date information.
E2B
Using MCP to run code via e2b.
Neon Database
MCP server for interacting with Neon Management API and databases
Exa Search
A Model Context Protocol (MCP) server lets AI assistants like Claude use the Exa AI Search API for web searches. This setup allows AI models to get real-time web information in a safe and controlled way.
Qdrant Server
This repository is an example of how to create a MCP server for Qdrant, a vector search engine.