{"id":50196572,"url":"https://github.com/timoncool/ace-step-studio","last_synced_at":"2026-05-25T18:04:43.182Z","repository":{"id":350573490,"uuid":"1207422559","full_name":"timoncool/ACE-Step-Studio","owner":"timoncool","description":"Portable AI music generator — full songs with vocals, covers, music videos. One-click install, 100% offline, NVIDIA GPU.","archived":false,"fork":false,"pushed_at":"2026-05-04T05:50:21.000Z","size":31692,"stargazers_count":145,"open_issues_count":10,"forks_count":11,"subscribers_count":5,"default_branch":"master","last_synced_at":"2026-05-04T07:34:03.399Z","etag":null,"topics":["ace-step","ai-music","gradio","local-ai","lyrics","music-generation","music-video","offline","one-click-install","portable","pytorch","text-to-music","vocals","windows"],"latest_commit_sha":null,"homepage":"https://github.com/timoncool/ACE-Step-Studio/releases","language":"Python","has_issues":true,"has_wiki":null,"has_pages":null,"mirror_url":null,"source_name":null,"license":"mit","status":null,"scm":"git","pull_requests_enabled":true,"icon_url":"https://github.com/timoncool.png","metadata":{"files":{"readme":"README.md","changelog":"CHANGELOG.md","contributing":null,"funding":".github/FUNDING.yml","license":"LICENSE","code_of_conduct":null,"threat_model":null,"audit":null,"citation":null,"codeowners":null,"security":null,"support":null,"governance":null,"roadmap":null,"authors":null,"dei":null,"publiccode":null,"codemeta":null,"zenodo":null,"notice":null,"maintainers":null,"copyright":null,"agents":"AGENTS.md","dco":null,"cla":null},"funding":{"custom":["https://github.com/timoncool/ACE-Step-Studio/blob/master/DONATE.md","https://dalink.to/nerual_dreming","https://boosty.to/neuro_art"]}},"created_at":"2026-04-10T23:34:01.000Z","updated_at":"2026-05-04T05:50:26.000Z","dependencies_parsed_at":null,"dependency_job_id":null,"html_url":"https://github.com/timoncool/ACE-Step-Studio","commit_stats":null,"previous_names":["timoncool/ace-step-studio"],"tags_count":1,"template":false,"template_full_name":null,"purl":"pkg:github/timoncool/ACE-Step-Studio","repository_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/timoncool%2FACE-Step-Studio","tags_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/timoncool%2FACE-Step-Studio/tags","releases_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/timoncool%2FACE-Step-Studio/releases","manifests_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/timoncool%2FACE-Step-Studio/manifests","owner_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/owners/timoncool","download_url":"https://codeload.github.com/timoncool/ACE-Step-Studio/tar.gz/refs/heads/master","sbom_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/timoncool%2FACE-Step-Studio/sbom","scorecard":null,"host":{"name":"GitHub","url":"https://github.com","kind":"github","repositories_count":286080680,"owners_count":33486790,"icon_url":"https://github.com/github.png","version":null,"created_at":"2022-05-30T11:31:42.601Z","updated_at":"2026-05-25T14:31:05.219Z","status":"ssl_error","status_checked_at":"2026-05-25T14:31:02.878Z","response_time":57,"last_error":"SSL_read: unexpected eof while reading","robots_txt_status":"success","robots_txt_updated_at":"2025-07-24T06:49:26.215Z","robots_txt_url":"https://github.com/robots.txt","online":false,"can_crawl_api":true,"host_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub","repositories_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories","repository_names_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repository_names","owners_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/owners"}},"keywords":["ace-step","ai-music","gradio","local-ai","lyrics","music-generation","music-video","offline","one-click-install","portable","pytorch","text-to-music","vocals","windows"],"created_at":"2026-05-25T18:04:27.600Z","updated_at":"2026-05-25T18:04:43.169Z","avatar_url":"https://github.com/timoncool.png","language":"Python","funding_links":["https://github.com/timoncool/ACE-Step-Studio/blob/master/DONATE.md","https://dalink.to/nerual_dreming","https://boosty.to/neuro_art"],"categories":[],"sub_categories":[],"readme":"\u003cdiv align=\"center\"\u003e\n\n# ACE-Step Studio\n\n**Suno at home. Local AI music generation studio — songs, vocals, lyrics, covers, music videos.**\n\n[![Stars](https://img.shields.io/github/stars/timoncool/ACE-Step-Studio?style=flat-square)](https://github.com/timoncool/ACE-Step-Studio/stargazers)\n[![License](https://img.shields.io/github/license/timoncool/ACE-Step-Studio?style=flat-square)](LICENSE)\n[![Last Commit](https://img.shields.io/github/last-commit/timoncool/ACE-Step-Studio?style=flat-square)](https://github.com/timoncool/ACE-Step-Studio/commits/master)\n[![Downloads](https://img.shields.io/github/downloads/timoncool/ACE-Step-Studio/total?style=flat-square)](https://github.com/timoncool/ACE-Step-Studio/releases)\n\n**[Русская версия](README_RU.md)**\n\n![ACE-Step Studio](docs/screenshots/main-ui.png)\n\n\u003c/div\u003e\n\nCreate full songs with vocals, lyrics, covers, remixes and music videos — **100% local**, no cloud, no subscriptions, no internet required. One-click install on Windows, runs on any NVIDIA GPU with 12+ GB VRAM.\n\nBuilt on [ACE-Step 1.5 XL](https://github.com/ace-step/ACE-Step-1.5) — the open-source 4B parameter DiT music generation model.\n\n## Why ACE-Step Studio?\n\n- **Free forever** — no API keys, no credits, no usage limits\n- **Private** — your music never leaves your machine\n- **Portable** — everything in one folder, copy to USB, delete = uninstall\n- **One-click** — `install.bat` → `run.bat` → make music\n\n## Features\n\n### Music Generation\n- **Full songs with vocals** — up to 8 minutes, any language, any genre\n- **Simple \u0026 Custom modes** — describe what you want or fine-tune every parameter\n- **3 XL Models** — XL Turbo (8 steps, fast), XL SFT (50 steps, max quality), XL Turbo BF16 (compact, 7.5 GB)\n- **AI Lyrics \u0026 Style** — LLM generates lyrics and enriches style descriptions\n- **Hot Model Switching** — change DiT/LM models without restart\n- **Batch generation** — create multiple variations at once\n- **10 samplers, 7 schedulers** — euler, heun, midpoint, dopri5, deis, ipndm, and more\n- **LoRA support** — load LoRA weights at inference time\n- **ID3 tags** — MP3 files include title, artist, cover art, lyrics, BPM\n\n### Cloud LLM \u0026 Image (optional, off by default)\n- **OpenRouter for lyrics \u0026 style** — bring-your-own-key alternative to the local LM. Pick any model (Claude, GPT-4o, DeepSeek, Llama 3.x, etc.), get instant lyrics + caption + key/BPM/duration metadata without using GPU VRAM. Local LM keeps working in parallel — toggle anytime.\n- **Pollinations.ai cover generation** — auto-generate album covers in parallel with audio (server-side, fire-and-forget, never blocks audio gen). The visual prompt comes straight from the OpenRouter LLM (which writes a 1–2 sentence visual description tailored to the song's lyrics and mood) or from a keyword fallback. Anonymous tier works; bring your own token for higher rate limits and no watermark.\n- **Manual cover regen modal** — picture-with-pencil button on every track. Pick any Pollinations model, write your own prompt, \"Try again\" until you like it, **or upload your own image from disk** (JPEG/PNG/WEBP, ≤10MB). Saved cover replaces both `songs.cover_url` and the embedded ID3 frame inside the MP3, so external players see your picked image too.\n- **Independent toggles** — every cloud feature is opt-in. Use only Pollinations covers + local LM, or only OpenRouter lyrics + auto-picsum covers, or both, or neither. Local-only mode is the default.\n\n### Cover \u0026 Remix\n- **Cover mode** — transform existing audio into a new style while keeping the melody\n- **Repaint mode** — regenerate specific sections of a song (region selection on waveform)\n- **Reference audio** — use a reference track to guide the generation style\n- **Audio strength control** — blend between source and generated audio\n\n### Video Studio\n\n![Video Studio](docs/screenshots/video-studio.png)\n\n- **Music video generator** — NCS-style visualizers with 10 presets\n- **Karaoke lyrics** — synchronized LRC subtitles with 3 styles (lines, scroll, karaoke fill)\n- **WYSIWYG editor** — drag elements, scroll to resize, selection frames\n- **Aspect ratios** — 16:9, 9:16 (Reels/TikTok), 1:1 (Instagram)\n- **12 effects** — shake, glitch, VHS, CCTV, scanlines, bloom, film grain, strobe, vignette, hue shift, letterbox, pixelate\n- **Background** — random, custom image, Pexels search, video backgrounds\n- **Server-side rendering** — native ffmpeg with NVENC GPU acceleration\n\n### Audio Tools\n- **Audio editor** — trim, fade, effects (AudioMass)\n- **Stem extraction** — separate vocals, drums, bass, other (Demucs)\n- **LRC download** — export synchronized lyrics\n\n### Model Tools\n- **BF16 Converter** — convert safetensors from FP32/FP16 to BFloat16 (~50% size reduction)\n- **Model Merger** — merge two ACE-Step models with adjustable alpha (3 methods)\n- **Bake LoRA** — bake LoRA weights into base model\n\n### Interface\n- **Single terminal** — one `run.bat`, Express manages Python/Gradio automatically\n- **Portable** — everything in one folder, no system-wide installs\n- **5 languages** — English, Russian, Chinese, Japanese, Korean\n- **LAN access** — use from any device on your network (phone, tablet)\n- **GPU monitoring** — live VRAM, RAM, CPU, temperature stats\n- **Dark/Light theme**\n\n## System Requirements\n\n| Component | Minimum | Recommended |\n|-----------|---------|-------------|\n| GPU VRAM | 12 GB | 20+ GB |\n| RAM | 16 GB | 32 GB |\n| Disk | 30 GB | 60 GB (all models) |\n| OS | Windows 10/11 | Windows 11 |\n| GPU | RTX 3060+ | RTX 4090 |\n\n## Quick Start\n\n\u003e 🚀 **One-click cross-platform install via Pinokio:** [![Install on Pinokio](https://img.shields.io/badge/Install_on-Pinokio-7c3aed?style=flat-square)](https://pinokio.co/item?uri=https://github.com/timoncool/ACE-Step-Studio-pinokio) [![Open in Pinokio](https://img.shields.io/badge/Open_in-Pinokio-6d28d9?style=flat-square)](https://beta.pinokio.co/apps/github-com-timoncool-ace-step-studio-pinokio)\n\u003e\n\u003e Works on Windows / Linux (x64 \u0026 aarch64) / macOS (Apple Silicon \u0026 Intel). No `install.bat` required — Pinokio bundles Python, Node, ffmpeg, venv and picks the right PyTorch build for your GPU automatically.\n\u003e\n\u003e Pinokio launcher repo: **[timoncool/ACE-Step-Studio-pinokio](https://github.com/timoncool/ACE-Step-Studio-pinokio)**\n\n---\n\nOr install manually on Windows:\n\n### 1. Clone\n\n```bash\ngit clone https://github.com/timoncool/ACE-Step-Studio.git\ncd ACE-Step-Studio\n```\n\n### 2. Install\n\n```\ninstall.bat\n```\n\nSelect your GPU type (CUDA 12.8 / 12.6 / 12.4). Installs portable Python 3.12, PyTorch, Node.js 22, and all dependencies — nothing system-wide.\n\n### 3. Run\n\n```\nrun.bat\n```\n\nBrowser opens automatically at http://localhost:3001. Models download on first run (~7.5 GB for default BF16 model).\n\n## Launchers\n\n| Script | Description |\n|--------|-------------|\n| `run.bat` | Standard launch — DiT + LM (0.6B PT), full features |\n| `run-no-lm.bat` | Launch without LM — more VRAM for DiT, cover/repaint work, no AI lyrics/thinking |\n| `run-dev.bat` | Dev mode — 3 terminals with Vite HMR |\n| `install.bat` | One-click installer |\n| `update.bat` | Update code + deps + rebuild frontend |\n| `reinstall.bat` | Clean reinstall (preserves models and data) |\n| `download_model.bat` | Pre-download models |\n\n## Models\n\n| Model | Size | Steps | Speed | Quality |\n|-------|------|-------|-------|---------|\n| XL Turbo BF16 | 7.5 GB | 8 | Fast | High |\n| XL Turbo | 18.8 GB | 8 | Fast | Very High |\n| XL SFT | 18.8 GB | 50 | Slow | Highest |\n| XL Merge SFT+Turbo | 18.8 GB | 12 | Medium | Very High |\n\n### LM Models (text/lyrics AI)\n\n| Model | VRAM | Quality |\n|-------|------|---------|\n| 0.6B | ~0.5 GB | Basic |\n| 1.7B | ~1.5 GB | Good |\n| 4B | ~4 GB | Best |\n\nLM backend: **PT** (PyTorch, lighter) or **vLLM** (faster inference, more VRAM).\n\n## API Keys (optional)\n\nACE-Step Studio is fully usable **without any API keys** — local DiT + local LM cover everything music-generation related. The keys below unlock optional cloud services that some users prefer for convenience or quality. They are stored in browser `localStorage` only, never sent to any server but the provider's own.\n\n\u003e **TL;DR — both providers can be used 100 % free.**\n\u003e OpenRouter has dozens of completely free models you can pick (DeepSeek R1 free, Llama 3.3 70B free, Gemini 2.0 Flash free, Qwen 2.5 free, Mistral Small free…) — just create a key and choose any model with a `:free` tag.\n\u003e Pollinations.ai works **without any account at all** on the anonymous tier (slower, occasional watermark on some models) — leave the key field blank and it just works.\n\n| Provider | What it does in the app | Where to get it | Free tier |\n|---|---|---|---|\n| **OpenRouter** | Generates lyrics + caption + BPM/key/duration metadata + a visual cover prompt from your one-line description (replaces the local LM). Lets you pick Claude / GPT-4o / DeepSeek / Llama / Mistral / Gemini / any of 200+ models. | [openrouter.ai/keys](https://openrouter.ai/keys) — sign in with Google/GitHub, click *Create Key*. | **Yes — many fully free models** (filter the model picker by `:free`): DeepSeek R1 free, Llama 3.3 70B Instruct free, Gemini 2.0 Flash free, Qwen 2.5 free, Mistral Small 3 free, and more. Paid models are pay-per-token from your wallet — no monthly subscription required. |\n| **Pollinations.ai** | Generates the album cover image in parallel with audio gen, plus powers the manual cover-regen modal. Token also unlocks the full image-model catalogue (FLUX, Qwen-Image, Klein, GPT-Image, Z-Image, …) and removes the watermark. | [auth.pollinations.ai](https://auth.pollinations.ai) — sign in, copy `pk_…` (public) or `sk_…` (private) key. | **Yes — fully free**, anonymous tier works without any account or key. Slower (1 req/15 s) and may have a small watermark on certain models. With a free token: 1 req/5 s + no watermark + full model list. |\n\n### Where to enter them\n\n- **OpenRouter** → Create panel → Advanced → toggle *\"Use OpenRouter\"* → paste key, pick model, *Test*.\n- **Pollinations** → Create panel → Advanced → *Cover image (Pollinations.ai)* → toggle *\"Generate covers via Pollinations.ai\"* → paste key (optional), pick model, *Test*.\n\nBoth toggles persist across sessions and are independent — turn either one off to fall back to the local pipeline (LM for lyrics, picsum for covers).\n\n### How private is this?\n\n- Keys live only in your browser's `localStorage` for this site.\n- They are sent **only** to `openrouter.ai` / `gen.pollinations.ai` over HTTPS, attached to that single API call. ACE-Step Studio does not have a backend account, telemetry, or proxy server.\n- Cover images you generate are written to `app/server/public/audio/\u003cuserId\u003e/covers/\u003csongId\u003e.jpg` on your machine. Nothing is uploaded anywhere.\n- If you don't want any cloud calls, simply leave both toggles off — the entire app works offline.\n\n## Architecture\n\n```\nACE-Step-Studio/\n├── app/              # React + Express frontend \u0026 backend\n├── ACE-Step-1.5/     # Python ML pipeline\n├── python/           # Portable Python 3.12 (created by install.bat)\n├── node/             # Portable Node.js 22 (created by install.bat)\n├── models/           # HuggingFace cache (created at runtime)\n├── run.bat           # Standard launcher\n├── run-no-lm.bat     # Launch without LM\n├── install.bat       # One-click installer\n├── update.bat        # Updater\n└── CHANGELOG.md      # Version history\n```\n\n## Updating\n\n```\nupdate.bat\n```\n\nPulls latest code, updates Python/Node deps, rebuilds frontend.\n\n## Contributing\n\nContributions welcome! Here's how to help:\n\n- **Report bugs** — [open an issue](https://github.com/timoncool/ACE-Step-Studio/issues)\n- **Suggest features** — [start a discussion](https://github.com/timoncool/ACE-Step-Studio/issues)\n- **Submit PRs** — see [AGENTS.md](AGENTS.md) for architecture, coding conventions, and pitfalls\n\nAreas where help is especially needed:\n- macOS / Linux support\n- New visualizer presets for Video Studio\n- Translations (i18n)\n- LoRA training UI improvements\n- Documentation \u0026 tutorials\n\n## Other Portable Neural Networks\n\n| Project | Description |\n|---------|-------------|\n| [Foundation Music Lab](https://github.com/timoncool/Foundation-Music-Lab) | Music generation + timeline editor |\n| [VibeVoice ASR](https://github.com/timoncool/VibeVoice_ASR_portable_ru) | Speech recognition (ASR) |\n| [LavaSR](https://github.com/timoncool/LavaSR_portable_ru) | Audio quality enhancement |\n| [Qwen3-TTS](https://github.com/timoncool/Qwen3-TTS_portable_rus) | Text-to-speech by Qwen |\n| [SuperCaption Qwen3-VL](https://github.com/timoncool/SuperCaption_Qwen3-VL) | Image captioning |\n| [VideoSOS](https://github.com/timoncool/videosos) | AI video production |\n| [RC Stable Audio Tools](https://github.com/timoncool/RC-stable-audio-tools-portable) | Music and audio generation |\n\n## Authors\n\n- **Nerual Dreming** — [Telegram](https://t.me/nerual_dreming) | [neuro-cartel.com](https://neuro-cartel.com) | [ArtGeneration.me](https://artgeneration.me)\n- **Neiro-Soft** — [Telegram](https://t.me/neuroport) | portable neural network builds\n\n## Acknowledgments\n\n- **[ACE-Step Team](https://github.com/ace-step)** — open source ACE-Step 1.5 music generation model\n- **[fspecii](https://github.com/fspecii/ace-step-ui)** — original ACE-Step UI\n- [AudioMass](https://audiomass.co/) — browser audio editor\n- [Demucs](https://github.com/facebookresearch/demucs) — stem extraction by Meta\n- [Pexels](https://www.pexels.com/) — free stock photos/videos\n- [Gradio](https://gradio.app/) — ML model serving\n- [FFmpeg](https://ffmpeg.org/) — video encoding\n\n## Support This Project\n\nI build software and do research in AI and music generation. Most of what I create is free and open source. Your donations allow me to keep creating and exploring without worrying about where the next meal comes from =)\n\n**[All donation methods](DONATE.md)** | **[dalink.to/nerual_dreming](https://dalink.to/nerual_dreming)** | **[boosty.to/neuro_art](https://boosty.to/neuro_art)**\n\n- **BTC:** `1E7dHL22RpyhJGVpcvKdbyZgksSYkYeEBC`\n- **ETH (ERC20):** `0xb5db65adf478983186d4897ba92fe2c25c594a0c`\n- **USDT (TRC20):** `TQST9Lp2TjK6FiVkn4fwfGUee7NmkxEE7C`\n\n---\n\n## Star History\n\n\u003ca href=\"https://www.star-history.com/?repos=timoncool%2FACE-Step-Studio\u0026type=date\u0026legend=top-left\"\u003e\n \u003cpicture\u003e\n   \u003csource media=\"(prefers-color-scheme: dark)\" srcset=\"https://api.star-history.com/svg?repos=timoncool/ACE-Step-Studio\u0026type=date\u0026theme=dark\u0026legend=top-left\" /\u003e\n   \u003csource media=\"(prefers-color-scheme: light)\" srcset=\"https://api.star-history.com/svg?repos=timoncool/ACE-Step-Studio\u0026type=date\u0026legend=top-left\" /\u003e\n   \u003cimg alt=\"Star History Chart\" src=\"https://api.star-history.com/svg?repos=timoncool/ACE-Step-Studio\u0026type=date\u0026legend=top-left\" /\u003e\n \u003c/picture\u003e\n\u003c/a\u003e\n","project_url":"https://awesome.ecosyste.ms/api/v1/projects/github.com%2Ftimoncool%2Face-step-studio","html_url":"https://awesome.ecosyste.ms/projects/github.com%2Ftimoncool%2Face-step-studio","lists_url":"https://awesome.ecosyste.ms/api/v1/projects/github.com%2Ftimoncool%2Face-step-studio/lists"}