iArt.ai
AI 모션 그래픽 소프트웨어, 누구나 사용할 수 있는 After Effects 대안, 텍스트 프롬프트로 몇 분 안에 애니메이션·설명 영상 제작
아직 감지된 업데이트가 없습니다. GitHub 릴리스가 나오면 여기에 보입니다.
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 확인하기메이커가 쓴 소개는 위의 한 줄이 전부입니다. 더 자세한 내용은 저장소 README에서 읽을 수 있습니다. 저장소 README ↗
README 전체를 불러오는 중입니다.
README 전문 보기 ↗Free online MP3 tools built with Next.js (App Router). All audio is processed inside the visitor's browser — files are never uploaded to a server.
| Route | Tool |
|---|---|
/ | MP3 converter for any audio/video file, paste-a-link downloader and voice recorder |
/youtube-to-mp3, /youtube-to-mp4, /online-video-downloader | Paste a link → download MP4 or convert to MP3 (needs the download service) |
/mp4-to-mp3, /video-to-mp3, /wav-to-mp3, /m4a-to-mp3, /flac-to-mp3, /aac-to-mp3, /ogg-to-mp3, /opus-to-mp3, /webm-to-mp3, /mov-to-mp3, /mkv-to-mp3, /podcast-to-mp3 | Format-specific converter pages |
/voice-to-mp3 | Microphone recorder → MP3 |
/mp3-cutter | Waveform trimmer: keep/remove selection, fade in/out, preview |
/mp3-compressor | Re-encode at 32–160 kbps, shows size saved |
/audio-joiner | Merge files in any order with optional gaps |
/video-to-mp3, /audio-to-mp3, /link-to-mp3, /mp3-tools | Hub pages that link every tool of a sidebar group |
/transcribe-audio-to-text | Speech to text with Whisper in the browser: language detection, TXT/SRT/VTT export |
/about, /contact, /privacy-policy, /terms, /copyright | Info and legal pages |
lib/audio-engine.js)public/ffmpeg-worker.js + @ffmpeg/core). The ~32 MB core is downloaded only when needed.public/mp3-worker.js).Files are processed one at a time to keep memory use low; the limit is 500 MB per file.
public/transcribe-worker.js)The transcriber runs OpenAI's Whisper model (MIT) with transformers.js
(Apache-2.0, loaded from jsDelivr at a pinned version) in a module worker. Audio is decoded with the engine above,
resampled to 16 kHz mono and never leaves the device. The browser downloads, on first use, the library and the ONNX runtime from
jsDelivr and the chosen model (Xenova/whisper-tiny, -base or -small, 8-bit) from Hugging Face, and caches them.
transformers.js has no language detection, so the worker scores Whisper's language tokens on the first 30 seconds.
With WebGPU (desktop Chrome/Edge, recent Safari) it runs on the graphics card using the onnx-community/whisper-*
models (fp32 encoder, 4-bit decoder); otherwise, or if the GPU fails to load or run, it falls back to the 8-bit
WebAssembly models. The model starts downloading as soon as a file is chosen or a recording starts.
To update transformers.js, change the version in the worker's import URL.
Browsers cannot fetch videos from other websites, so link downloads use a small service in
downloader/ (Node + yt-dlp + FFmpeg, Docker, automatic HTTPS with Caddy).
Run it on a VPS, then set this environment variable in Vercel and redeploy:
NEXT_PUBLIC_DOWNLOADER_URL=https://dl.mp3convert.tools
Without it, the "Paste Link" tools show a friendly "not set up yet" message and everything else keeps working.
npm install # also copies the FFmpeg core into public/ffmpeg (postinstall)
npm run dev # http://localhost:3000
npm run build && npm start
public/ffmpeg/ is generated by scripts/copy-ffmpeg.mjs and is not committed. Hosts that run npm install
(Vercel, Netlify, etc.) create it automatically.
lib/tools.jslib/site.js. POPULAR_NAV is the always-open list and
NAV_GROUPS the groups. A group with a hub links its heading to that hub page, which shows a card for every tool
of the group (the hub field of the page in lib/tools.js), and its tools get a Home › Hub › Tool breadcrumb.
Sidebar labels are short (nav in lib/i18n/ui-en.js); the full keyword lives in each page's title and H1.The site is available in English (at the root, e.g. /youtube-to-mp3) and in Spanish, Portuguese, French, German,
Russian, Turkish, Arabic (right-to-left), Hindi and Indonesian under a language prefix (/es/youtube-to-mp3).
proxy.js serves English from the root; pages live in app/[lang].
lib/tools.js, the homepage in lib/content/home.js, the About/Contact/legal pages in
lib/content/pages.js, and interface text (buttons, menus, messages) in lib/i18n/ui-en.js.lib/i18n/locales/<lang>/{ui,home,pages,tools}.json, with the same structure as the English source.
Anything missing falls back to English. A language without translation files is hidden from the switcher, the
sitemap and hreflang tags, and is marked noindex.lib/i18n/config.js (keep proxy.js in sync when adding a language).