iArt.ai
AI 모션 그래픽 소프트웨어, 누구나 사용할 수 있는 After Effects 대안, 텍스트 프롬프트로 몇 분 안에 애니메이션·설명 영상 제작
개인 계정의 공개 저장소 소유자는 GitHub로 로그인해 운영자를 확인하고 소개를 직접 관리할 수 있습니다. 관리자 승인은 필요하지 않습니다.
로그인한 뒤 확인 파일을 만들어 이 저장소의 기본 브랜치에 30분 안에 커밋해 주세요. 다른 브랜치나 포크의 파일은 인정하지 않습니다. 파일을 올린 뒤 아래에서 확인하면 관리자 승인 없이 관리할 수 있습니다. 자격을 유지하려면 파일을 남겨 두세요.
로그인이 필요하면 로그인 화면으로 이동합니다. 로그인 후 이곳으로 돌아와 원하는 확인 버튼을 다시 눌러 주세요.
프로젝트 폴더에서 /nomorevibe verify를 실행하면 소유권을 확인하고 소개를 직접 관리할 수 있습니다.
1분 만에 확인하기Free, open-source guitar amp sim and multi-FX with neural amp models (NAM). Mac and Windows
저장소 README에서 가져옴
README 전체를 불러오는 중입니다. 아래는 저장된 발췌입니다.
A guitar amp simulator and multi-effects processor in C++ on JUCE, with neural amp captures (NAM) running on NeuralAmpModelerCore, the official NAM engine. Built as a personal tool for practicing and writing, and free software: AGPL-3.0-or-later (see License).
Current target: a standalone macOS app that plays live through a Focusrite Scarlett Solo 4th Gen. VST3/AU plugin export comes later from the same JUCE target.
Everything in the plan is built (see docs/PROGRESS.md for the measurements behind each item, and docs/ASSUMPTIONS.md for the choices waiting on Sean's review):README 전문 보기 ↗A guitar amp simulator and multi-effects processor in C++ on JUCE, with neural amp captures (NAM) running on NeuralAmpModelerCore, the official NAM engine. Built as a personal tool for practicing and writing, and free software: AGPL-3.0-or-later (see License).
Current target: a standalone macOS app that plays live through a Focusrite Scarlett Solo 4th Gen. VST3/AU plugin export comes later from the same JUCE target.
Everything in the plan is built (see docs/PROGRESS.md for the measurements behind each item, and docs/ASSUMPTIONS.md for the choices waiting on Sean's review):
.nam capture as a ninth, picked from the stage's shelf of minis or a MIDI footswitch (program change 1 to 9). Only the selected amp runs: a switch warms the incoming amp up unheard, then crossfades, so it never clicks and lands about 0.1 s later. Each amp's Gain blends five captures across the knob. The built-in captures are trained from the project's own amp models, not captures of real amps.The GUI (docs/UI_DESIGN.md, DESIGN.md) is one main screen on a fixed 1280 x 760 canvas that scales with the window, every pixel drawn in code:
[ and ] step the tabs.Keys: Left and Right step presets, 1 to 8 recall scenes, T opens the tuner, Escape closes an overlay, Cmd-Z and Shift-Cmd-Z undo and redo.
You need Apple's command line tools (xcode-select --install) and CMake (installed in ~/.local/bin). Full setup, including the Scarlett Solo gotchas, is in docs/SETUP.md.
git submodule update --init --recursive # JUCE and NeuralAmpModelerCore
cmake -B build -DCMAKE_BUILD_TYPE=Release # configure
cmake --build build -j # compile
ctest --test-dir build --output-on-failure # run the tests
open build/BellyDSP_artefacts/Release/Standalone/BellyDSP.app
Releases are published on the public repo, https://github.com/SeanSnaider/BellyDSP/releases, and the app updates itself (Sparkle on the Mac, WinSparkle on Windows). Every release is built from a commit that's on the public repo's main and tagged there, so the exact source of every download is public. One command checks that, builds, tests, signs, packages, tags, pushes the tag, and publishes: tools/release/release.sh <version>. docs/RELEASING.md has the one-time setup and the details; docs/INSTALL.md is the install guide.
| Path | What it is |
|---|---|
src/dsp/ | The DSP: the Block interface, the chain, every amp, cab, filter, and effect |
src/PluginProcessor.* | Parameters, state, file loading, MIDI, and the audio callback, which just drives the chain |
src/BlockParameters.*, src/Presets.* | Parameter glue for each effect block, and the preset format |
src/PluginEditor.* | The editor: lays out the GUI and runs presets, scenes, MIDI learn menus, and undo |
src/platform/ | What the shipped app knows about itself: its version, its bundled files, and the auto-updater |
src/ui/ | The GUI: tokens and fonts, LookAndFeel, controls, meters, the top bar, the stage, the board and its pedal art, the amp heads, the cab, every block's inspector, the scenes, and the overlays (tuner, Match tone, Settings) |
presets/factory/ | The factory presets, embedded in the app |
docs/ui/amp-ui-handoff/ | The original GUI handoff, retired as a source of truth by the redesign (docs/UI_DESIGN.md) |
resources/fonts/ | Geist and Fraunces (SIL OFL, licences alongside), embedded in the app |
tools/render/ | ampsim_render: runs a WAV through the chain offline, with CPU timing |
tools/device_probe/ | ampsim_device_probe: opens the interface and reports timing and input levels |
tools/capture/, tools/train_capture.sh, tools/fetch_nam_input.sh | Capturing your own gear for NAM: ampsim_capture plays NAM's input file through the gear and records it, and the trainer wrapper runs NAM's official trainer on it (docs/CAPTURING.md) |
tools/content/ | The scripts that imported the bundled cab IRs and ranked them |
tools/release/, tools/fetch_deps.sh, tools/notices/ | The release pipeline: signing, packaging, appcasts, publishing, the update test, the licence notices |
installer/ | The DMG's read-me and the Windows installer script (Inno Setup) |
No .nam models are bundled. Download amp-only captures (no cab baked in) from TONE3000, or train your own with NAM's trainer, and keep them in models/, which is gitignored. NAM core's example models in third_party/NeuralAmpModelerCore/example_models/ work for testing.
21 public-domain (CC0) IRs from Jester Dyne Productions are bundled (content/, listed under "Built in" in the cab's library on the stage). Load your own .wav files too (free and paid packs are widely available); keep them in irs/, which is gitignored.
BellyDSP Copyright (C) 2026 Sean Snaider
This program is free software: you can redistribute it and/or modify it under the terms of the GNU Affero General Public License as published by the Free Software Foundation, either version 3 of the License, or (at your option) any later version (SPDX: AGPL-3.0-or-later). This program comes with ABSOLUTELY NO WARRANTY; without even the implied warranty of MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See LICENSE for the full text.
Source: https://github.com/SeanSnaider/BellyDSP. The app's brand menu links to the exact source of each version.
tools/spdx_headers.py).content/, as they're added) are licensed under CC BY 4.0, attribution "Sean Snaider". The bundled cab IRs are CC0 1.0.THIRD_PARTY_NOTICES.txt (generated by tools/notices/make_notices.cmake, shipped in the app and the downloads).content/ | Captures and IRs bundled with the app (so far two CC0 cab IR packs, 21 IRs), each with its licence in manifest.json |
release-notes/ | One Markdown file per released version |
.github/workflows/ | The Windows build and installer, the macOS tests, the website's deploy, and the daily download stats |
site/, tools/site/ | The website, bellydsp.com: plain HTML and CSS, the script that writes its download data, and its checks (docs/WEBSITE.md) |
tests/ | The test suite (ampsim_tests), including real-time safety and NAM differential tests |
third_party/ | JUCE 8.0.15 and NeuralAmpModelerCore v0.6.0, as pinned git submodules |
prototypes/ | Python lab bench. Algorithms get prototyped and listened to here before being ported |
docs/BUILD_PLAN.md | Full feature spec, architecture, build order, and decision log |
docs/PROGRESS.md | Phase-by-phase status |
docs/SETUP.md | Mac + Scarlett Solo setup and troubleshooting |
docs/DSP_REFERENCE.md | Glossary and the sources behind each algorithm |
docs/ASSUMPTIONS.md | Every decision made on Sean's behalf during the unattended build, to review |
docs/NAM_UPSTREAM.md | NAM core findings worth taking upstream |
docs/RELEASING.md, docs/INSTALL.md | How releases and updates work (for Sean), and how to install (for players) |
docs/WEBSITE.md | How the website is built, deployed, and counted, and the domain's one-time setup |
CLAUDE.md | Context and rules for AI assistants working in this repo |
LICENSE | The GNU AGPL v3, the licence of all the code |
tools/spdx_headers.py | Adds (or checks) the SPDX licence header on every source file |