PyWwise active

fmod-studio-mcp

FMOD

fmod-studio-mcp is an MCP (Model Context Protocol) server that connects to FMOD Studio's built-in scripting terminal over TCP and exposes the full FMOD Studio Scripting API as discrete, schema-validated MCP tools. Instead of editing project XML on disk, it talks to the running editor so changes appear immediately without requiring a project reload. The tool set is auto-generated from the crawled FMOD Scripting API reference — roughly 148 named tools covering events, tracks, banks, sounds, assets, and system operations — plus a set of generic property/relationship tools for dynamic, schema-defined members the static docs don't enumerate. A composite tool (`fmod_create_event`) handles the common one-shot workflow of creating an event, importing audio, wiring a track and instrument, and routing to a bank in a single call. Designed for use with AI coding agents, it requires FMOD Studio 2.02+ with the scripting console active (Ctrl+0) and Python 3.10+. The spec can be regenerated for new FMOD versions by re-crawling the official reference. All edits affect the live project, so version control and explicit saves via `fmod_project_save` are strongly recommended.

Reaper-Waapi-Transfer

Wwise

A Reaper extension that streamlines the workflow between Reaper DAW and Wwise middleware. It automates the process of transferring rendered audio files from Reaper sessions into Wwise projects via the Wwise Authoring API (WAAPI). The extension eliminates manual copy-paste workflows when moving audio assets from DAW to middleware. While no longer actively maintained, it provides a foundation for automated asset pipeline integration between these tools.

SK Wwise MCP

Wwise

A modular suite of 12 MCP servers exposing 97 tools for Audiokinetic Wwise, built on the Wwise Authoring API (WAAPI). Each server handles a specific domain: browsing project hierarchies, creating and editing objects, managing audio import and SoundBank generation, controlling transport playback, querying profiler data, and automating UI workflows. The architecture uses a thread-safe WAAPI dispatcher with queue-based serialization to prevent race conditions. Ships as a standalone Windows executable requiring no Python installation, with pre-configured Agent Skills routing for multi-agent orchestration. Also includes WwiseConsole CLI integration for headless operations like project creation and migration. Designed to stay under per-server tool limits (15 tools each) to reduce LLM confusion, with 450 unit tests and 44 integration tests against live Wwise instances.

wwise-control

Wwise

wwise-control is a Claude Code skill and Python helper module that connects an AI agent to Wwise over the Wwise Authoring API (WAAPI). You describe what you want in plain language — creating containers, setting routing, adjusting properties, generating SoundBanks — and the agent translates that into pywwise calls without you writing or running scripts yourself. The package ships two files: a Python module (`wwise_for_claude.py`) that wraps pywwise with tested helpers for finding, creating, moving, renaming, and routing objects; and a SKILL.md that documents non-obvious WAAPI/WAQL quirks discovered through live testing (e.g. object identity requires `.guid` not `.id`, purely-numeric Bus names are rejected by Wwise). The skill instructs the agent to probe before writing, wrap batch edits in a Wwise undo group, verify results after each write, and confirm before deleting. Intended for use on version-controlled projects.