soundfont-renderer
Renders notes or MIDI files into realistic single-instrument audio tracks with selectable instruments and effects like reverb.
This server does not currently have a deployable release on Glama. Follow the repository's setup instructions to install and configure it for your MCP client, including any required credentials and local dependencies.
Soundfont Renderer(真乐器渲染)
Platform support: macOS — supported (developed & verified on this machine) · Windows — unverified (paths are cross-platform in the codebase, but no real-machine testing yet).
把音符或 MIDI 渲染成真乐器单轨:钢琴、小提琴、吉他、弦乐、铜管…… 引擎是 FluidSynth(LGPL-2.1,经 pyfluidsynth 动态调用),音色库是 GeneralUser GS v1.471(约 30MB,GM 全音色采样)。
中文说明 · License: AGPL-3.0-only
工具(MCP)
工具 | 输入 | 输出 |
| — |
|
|
|
|
|
| 同上 |
render_midi 内部先把 .mid 解析成和 render_notes 同一形状的音符列表,再走同一条渲染路径
(只有一个渲染规则)。
Related MCP server: Vibe Composer MIDI MCP
安装
引擎(libfluidsynth)与 Python 绑定装在插件自己的虚拟环境里,Micromamba 用户态安装,无需 sudo:
mkdir -p ~/Documents/ShadowRoom/_venvs && cd ~/Documents/ShadowRoom/_venvs
curl -sL https://github.com/mamba-org/micromamba-releases/releases/latest/download/micromamba-osx-arm64 -o micromamba && chmod +x micromamba
MAMBA_ROOT_PREFIX=$PWD/mamba-root ./micromamba create -y -p ./soundfont-renderer -c conda-forge python=3.12 fluidsynth numpy
./soundfont-renderer/bin/python -m pip install pyfluidsynthSoundFont 首次使用自动下载(约 30MB,sha256 校验,见下)。代理环境可设
SHADOWROOM_SOUNDFONT 指向自己下载的 .sf2 文件。
SoundFont:来源与许可证
来源:GeneralUser GS v1.471,作者 S. Christian Collins(http://www.schristiancollins.com), 镜像 JustEnoughLinuxOS/generaluser-gs。
许可证:允许在软件项目里自由使用与再分发(含商用音乐制作),但作者声明无法 100% 保证 其中全部采样的来源;全文见 docs/licenses.md 与 docs/GeneralUser-GS-LICENSE.txt。
因为这一条采样来源的不确定性,加上约 30MB 的二进制不适合放在 git 仓库里,本插件不随包携带 SoundFont,而是首次使用时下载到
~/Documents/ShadowRoom/_soundfonts/GeneralUser-GS-1.471.sf2并做 sha256 校验(f45b6b4a68b6bf3d792fcbb6d7de24dc701a0f89c5900a21ef3aaece993b839a)—— 与 stem-splitter 首次使用下载 Demucs 模型是同一个先例。FluidSynth 本体是 LGPL-2.1:本插件通过 pyfluidsynth(MIT)动态调用,未静态链接;以这种 方式分发不需要把插件改为 LGPL。
性能
SoundFont 在插件进程里只加载一次并常驻(进程存活期间所有渲染复用同一个 synth)。
离线渲染:音频按块从 synth 拉取,音符事件按采样位置触发,不受实时时钟限制—— 8 小节(15s 音频)渲染实测远低于 3 秒。
插件进程空闲 RSS < 150MB(GeneralUser GS 约 30MB 采样常驻后仍有富余,实测数字见验收)。
开发
PYTHONPATH=src python -m unittest discover -s tests -p "test_*_unittest.py"许可证
AGPL-3.0-only(见 LICENSE)。SoundFont 与 FluidSynth 各自的许可证见 docs/licenses.md。
Tool Schema Changelog
Recent tool additions, removals, and schema changes observed during successful MCP inspections.
Loading tool schema changelog…
This server cannot be deployed
Maintenance
Related MCP Connectors
Transcribe with speaker labels, split stems, remove noise, separate speakers, convert audio to MIDI.
Score video, make music as an agent: 100+ AI instruments play your MIDI, audio, stems, Logic/Ableton
Synth, sampler and music tracker for AI assistants: sound effects, UI sounds and songs as audio.
Browser DAW for AI assistants: build multitrack songs with MIDI, effects, stems and generation.
Related MCP Servers
- FlicenseNot gradedqualityDmaintenanceA MIDI composition system that enables AI assistants to create music through FluidSynth, with capabilities for playing notes, creating melodies, managing tracks, and exporting audio.1-
- AlicenseCqualityDmaintenanceEnables LLMs to compose and play multi-track MIDI music through natural language prompts. Supports outputting to software or hardware synthesizers for enhanced audio quality.225MIT
- AlicenseBqualityDmaintenanceEnables LLMs to compose music by describing musical intent using high-level operations like notes, chords, dynamics, and tempo, then renders the composition into standard MIDI files.47MIT
- AlicenseNot gradedqualityDmaintenanceEnables AI-assisted MIDI music generation and composition through natural language, supporting style, tempo, key, and instrument selection.MIT