New tool2026-10-18via GitHubVoice & Text-to-Speech

Audio Studio MCP

Name: Audio Studio MCP
Author: KanaWorks_AI (Tokyo AI Visuals LLC)

Audio Studio MCP is an open-source MCP server that lets Claude Code, Claude Desktop, Cursor and other MCP clients generate background music and sound effects locally with Stability AI's Stable Audio 3 model, then loop, normalize and export them.

Visit Audio Studio MCP ↗Repository ↗Release notes ↗

Free (open source)MIT (code); model weights under the Stability AI Community LicensePython (3.10-3.13, needs ffmpeg and git)★ 206 (as of 2026-10-19)by KanaWorks_AI (Tokyo AI Visuals LLC)

Why it matters

For people making videos, games or app demos who want custom background music and sound effects without a cloud service or per-clip fees. You describe the sound in plain language and your AI assistant generates and finishes the file on your own computer.

Highlights

Four tools: generate_audio (music, instrument loops or sound effects, 1-4 candidates per call), edit_audio (re-style, regenerate a time range, or extend a track), process_audio (trim, seamless loops, fades, loudness, WAV/OGG/FLAC export) and unload_models.
Runs Stable Audio 3 locally: MLX on Apple Silicon Macs and PyTorch on Windows PCs with an NVIDIA GPU (Windows support arrived in v0.2.0). Nothing is uploaded and no API key is needed.
Every file gets a JSON sidecar with its prompt, seed and settings, so results can be reproduced.
The server passes the official Stable Audio 3 prompt guide to the client, so casual requests are turned into prompts the model understands.
The maker reports about 10 s for 20 s of music on an M1 Mac and about 1.3 s on an RTX 3060 once the model is loaded.unverified

Taken from the project's own README, docs or launch page. Claims about speed or accuracy are the maker's unless stated otherwise.

Try it if: You make product videos, social clips or simple games and want royalty-style background music and sound effects generated on your own Mac or PC.

Worth knowing: Commercial use of the generated audio is governed by the Stability AI Community License, which the README says allows commercial use for individuals and organizations under US$1M in annual revenue; check the license yourself. The model download is about 6 GB on Mac and about 14 GB (with PyTorch) on Windows, and the Windows weights need a free Hugging Face account. It makes instrumental music only,

Quick facts

Pricing	Free; generation runs on your own hardware
Free or paid	Free (open source)
License	MIT (code); model weights under the Stability AI Community License
Language	Python (3.10-3.13, needs ffmpeg and git)
Platform	macOS on Apple Silicon (16 GB RAM recommended) or Windows 10/11 with an NVIDIA GPU (8 GB+ VRAM); Linux untested; Intel Macs and AMD GPUs not supported
GitHub stars	★ 206 (as of 2026-10-19)
Maker	KanaWorks_AI (Tokyo AI Visuals LLC)

Related ranking

See how established tools compare in Best Voice & Text-to-Speech. New releases aren't scored until we review them against the rubric.

What we couldn't verify

highlights[4] (performance figures are the maker's own)

Checked 2026-10-19 from the tool's own pages. Prices and star counts change, so confirm on the official site. Listing here is not a ranking or an endorsement.