
Apps
Announcements

🎵 OS Studio — A local AI songwriting studio Type an idea → Qwen writes the lyrics → YuE2 turns them into a full song → sing it in YOUR voice (record ~30 sec of yourself talking), or upload any voice. ✨ Guided 5-step composer, timed lyrics, AI-made styles, and a built-in player with Listen mode on your phone. Update: **⚙️ Windows or Linux · NVIDIA RTX 20-50 series, 8 GB+ · picks the right models for your card** Free & open source (GPL-3.0): https://github.com/TheAwaken1/os-studio

Ruach Studio 2.1.0-rc1: the Section editor, the Narrator in three columns, versions everywherePinokio's Update brings your Studio to 2.1.0-rc1, the first candidate for 2.1.0: everything made since 2.0.9, with 2.0.10's fixes in it. ✂️ The Section editor. Edit › Section editor… in a take's menu opens the song as a strip of its sections. Queue one edit a section (replace it, re-sing a line, sing it once more, insert an interlude or a copy of a part, repeat, remove, re-sing to the end, extend), then Assemble the take: one new take, the source kept as it is. 📖 The Narrator in three columns. The text, «At hand» and the voice side by side, with the lyrics editor's tools. Read the first 2, 3 or 5 minutes, or up to four versions at once. ✨ lets the Writer's model write one line on how to read the whole text, and Drama 3 hears it before every piece. 🗂️ Versions everywhere. The Writer, the Narrator and the Musician keep every version; a song lives in the Writer from its first save. 😀 Workspace emoji, history and MIDI. An emoji for each workspace and section; a history of plays and downloads (Ctrl+H, Ctrl+J); MIDI export in one window, from the score or per stem. 🎚️ Follow my sections is off by default now. It sits under the planning modes, says what it costs, and asks before a plan past 4:00. 🪟 Windows, driven for hours. Before this candidate we drove the Windows build on a clean Windows 11 machine with no graphics card for about four hours, round after round, until a round found nothing. The first start's install now shows its steps and takes your answer about the models, says how far the download came, and brings the same packages as on Linux; a machine with no NVIDIA card gets answers instead of errors; RuachStudio.exe starts the Studio without console windows, and the Weights panel works natively (issue #15). ⌨️ Pinokio takes your answer. Install, Update and More models now open their terminal to the keyboard, so «Download …? [Y/n]» can be answered. Thanks to @arkinou. A release candidate: what is in it works and was checked; the last polish comes in 2.1.0. Tell us what you hear. Release notes and downloads: https://github.com/igrbible/Ruach_Studio/releases/tag/v2.1.0-rc1 Everything, change by change: https://github.com/igrbible/Ruach_Studio/blob/main/CHANGELOG.md

PotionUI 0.3.0: the editors release with layers, region edits, one-click Upscale, Fix faces and Expand, and a video editor that saves your workPotionUI 0.3.0 is out, the editors release. Click Update in Pinokio to get it. This release is about editing what you made without leaving PotionUI. The Image editor saves your work and gets region edits, one-click tools and layers, the Video editor follows the same layout, and History shows what is running as it happens. What's new Saved edits: every Image and Video editor tab autosaves on the server. Open a new tab to pick up from your recent edits, and a picture saved from the editor keeps its steps, so Edit reopens it as you left it. Edit a region: mark an area with a brush or the new lasso, write what should go there, and any preset edits only that part. The result is blended back into the picture. Remove an object and Replace the background: one click, nothing to type. Expand: put a picture on a larger canvas in several formats at once, with social and banner sizes, your own size and a draggable position. Layers and paint: Layers and Edits are two panels, every paint layer is its own row with opacity, rename, reorder, Duplicate and Merge down, add a paint or picture layer any time, and one undo covers everything. Upscale by a factor or a size, and Fix faces and hands: tick the faces it found, add your own marks and set the strength. Video editor: the same rail, panels and options as the Image editor, clips saved as edits, and extend, upscale, restore, refine faces, colorize, day to night and remove object on the whole clip or a time range you mark. History shows running generations live: open one in Generation Details to watch it take shape, with its origin badge, and filter History by where a generation started. LTX-2.5 gets Text to Video, Image to Video and First to Last Frame. Kits: kit cards show every picture whole at its own aspect, and a new manage_kits tool lets MCP and chat create and fill kits. Nicer to use A menu bar, a floating action bar on the canvas, tool options as proper form fields, and Formulas and Settings in the dock. Choose which setup each editor task uses, and set your own default. A generated step whose picture changed underneath is marked so you can regenerate it. Kits, Library, History, Inspirations, Models, Prompts, Phrasebook and Auto-organize share one header with searchable filters. Reuse settings opens the Generate tab from History, prompts, model pages and the grid viewer. MiniMax-Music3 is faster, and lyrics are optional in Instrumental mode. Admin Activity shows the media tools queue. Community contributions Big thanks to M4-S4 for these fixes: On Windows, saving a downloaded model file retries when another program briefly holds it MiniMax-H3 crops reference pictures, clips and start pictures to fit instead of squeezing them Google Gemini shows the Veo 3.1 shutdown date, and its notes say what still works The Library records whether a video has sound, so tools that need sound are greyed out on a silent clip Replacing a Library item with a result in another format keeps the right file extension An updated plugin runs every module of its new version without a restart A plugin installed from another catalog shows that catalog's address Setup checks work for Video Director presets that have no text to video Fixed A paint stroke drawn right after adding a layer is no longer lost. The RunPod provider works with RunPod's new API. On Windows, web assets are served correctly even when the system maps them wrongly. A full-size Qwen-Image 2.1 edit no longer collides with models already in memory. Before you update Back up your data first. The first start moves your existing editor sessions and your default task setups to the new editor, and cannot be undone. Paint saved by an earlier version opens as a single layer. Windows is fully supported. Update Click Update in Pinokio, start PotionUI, then open the Image editor and try Expand on a picture. Full release notes: https://github.com/PotionUI/PotionUI/blob/master/CHANGELOG.md Website: https://potionui.com Discord: https://discord.gg/avR4trp3b8 Reddit: https://www.reddit.com/r/PotionUI/

Orange 1.2 is out — ComfyUI and cloud models in one appWhen I posted Orange 1.0, the idea was to turn a ComfyUI workflow into something you could hand to someone who just wanted to make an image. That’s still the idea. Since 1.0, I’ve added more local tools, made workflows easier to move between machines, and brought cloud generation into the same Generate page. More local workflows. The Curated Library now includes LTX 2.5 video with synchronized audio and an optional starting image, Qwen Image 2.1 Turbo Edit, and MOSS-Image for text-focused image generation. The redesigned Tool Editor makes optional inputs and workflow checks easier to configure. Orange can also reuse matching model files across different ComfyUI installations, even when the folders differ. Cloud image and video generation. You can build tools with Google, OpenAI, OpenRouter, Venice, xAI, Runway, and fal.ai. The new API filter in the Curated Library has ready-made Nano Banana 2.1 and GPT Image 2.6 Flare tools. Each can create an image from a prompt or edit one you attach, and the examples on the cards were generated by those models. Cloud jobs use the same Generate page, gallery, and downloads as local jobs. Outputs save locally, metadata removal is on by default, and API keys stay on the Orange server. You can use cloud tools alongside ComfyUI or choose Use cloud APIs only on a fresh Orange Only install. If you already have Orange in Pinokio, choose Update Orange and restart. For the cloud presets, open Admin → General Settings → API Keys, set up Google or OpenAI, then go to Tools → Curated Library → API. If you try it, I’d love to hear which workflows you’re putting in front of other people and where the setup still gets in your way. Orange v1.2.0 release notes · Pinokio launcher

HaruFeatured
Haru-4B voice conversations in your browser. Apple silicon Mac or NVIDIA CUDA on Windows/Linux. See README for GPU and system memory requirements.
cocktailpeanut/Haru.pinokiov8.0.0updated 7d ago
ZONOS3-MiniFeatured
Local multilingual speech generation, voice cloning, and voice steering with Zonos Voice Studio. https://github.com/Zyphra/ZONOS3-Mini
cocktailpeanut/zonos3-miniv8.0.0updated 5d ago
PotionUIFeatured
PotionUI: self-hosted AI image and video generation with a preset-driven interface, a native diffusers engine and ComfyUI support. Requires Windows or Linux with an NVIDIA GPU (driver 580 or newer, 8 GB VRAM and 16 GB RAM minimum). Other machines get a remote-only install.
PotionUI/potionui-pinokiov7.0updated 9d ago
FooocusPlusFeatured
An enhanced version of Fooocus giving you access to all of the latest AI image generation models
6Morpheus6/FooocusPlusv8.0updated 1mo ago
Drama 3 Offload StudioFeatured
Fish Audio Drama 3 (expressive TTS, voice cloning, voice design, directing notes) from the fishaudio/Drama-3 Space. Models: the full model for 24+ GB NVIDIA GPUs, the Small GGUF (drama3-small, Japanese and English) for 16 GB, and MLX 4-bit / 3-bit conversions of the full model for Apple Silicon (64 / 48 GB, untested on a Mac).
ThomasEricB/drama3-offload-studiov7.0updated 18h ago
YUE2 // GROOVEFeatured
YUE2 // GROOVE is the latest music studio built on the open-source Yue2 model and its inference stack. Generate high-quality full songs from style and lyrics with an editable score plan — powered by the latest YuE model — cover from audio with SheetSage2 and MERT2, refine and compare edits, and keep your works in a reusable, easy-to-manage library. You get high-quality creation with real creative control. Hardware: an NVIDIA GPU with 24 GB VRAM on Linux (YuE2's recommended setup, validated end-to-end on NVIDIA L4 hosts; a 16 GB memory budget runs everything the app can produce, and 12 GB runs the unquantized model at CFG 1.0 or for shorter songs) or an Apple Silicon Mac with 32 GB+ unified memory (where this app is developed and tested). Windows is best effort: install, launch, Cover and a full-length song verified on Windows 11 (RTX 2070, 8 GB); song generation there runs through the GGUF engine. Cards under 16 GB (and Windows) get the optional GGUF engine: the same model through yue2.cpp with an 8-bit backbone, 8.2 GB peak for a full song, measured indistinguishable from the reference rendering in a blind ABX, a different take for the same seed; the reference PyTorch configuration stays the default wherever it fits.
deadjoe/yue2-groove-pinokiov8.0updated 4d ago
SeedVR2-7B MLXFeatured
One-step SeedVR2-7B diffusion super-resolution upscaler, running natively on Apple Silicon via MLX. Choose between fp16 / int8 / sharp model variants, tune scale, softness, and seed, and compare before/after with a slider.
Rdx-ai-art/SeedVR2-MLX.pinokiov7.0updated 1mo ago
PulseForgeFeatured
A fully local, cross-platform audio visualizer editor. Create reactive music videos with layered graphics, AI-transcribed lyrics, and frame-perfect MP4 exports — all running in your browser
TheAwaken1/PulseForgev5.0updated 28d ago
theDAWFeatured
Full-featured DAW, DJ, and VJ app interoperable w/Ableton, Reaper, Resolume & more. Stable Audio 3, Magenta RT2, Lyria 3 Pro, Suno API, Chimera track fusion, Demucs stems, MIDI generate/notate, rule-based composer (harmony, counterpoint, fugue), orchestra sound bank w/concert-hall reverb, VST3 instruments, MPE, native Audima Labs Sway support, img > spectrogram > music, drawing > music, VST3 & .gan plugins, automix & key-lock, GLSL shaders, volumetric video, Quest 3 XR interface, MIDI auto-map, RAG assistant, asset library with example projects
gantasmo/theDAW-Pinokiov7.0updated 1mo ago
God's Eye ViewFeatured
A live 3D intelligence console for planet Earth.
bilawalsidhu/gods-eye-viewv3.6updated 5 min ago
MiniJamFeatured
Describe any song in plain English, compose a structured caption and lyrics locally, and generate it with MiniMax Music 3.
cocktailpeanut/minijamv7.0updated 2mo ago
MiniMax H3 ComfyUIFeatured
MiniMax H3 omni-modal video generation in ComfyUI. Text/image/video/audio in, video with native 32kHz stereo audio out (768p default, 1080p+ supported). Disk-optimized: pruned INT8 + NVFP4 weights (~63GB instead of ~290GB). NVIDIA only.
ThomasEricB/minimax-h3-pinokiov8.0updated 12d ago
SongGeneration StudioFeatured
AI Song Generation with Full Style Control - Generate complete songs with lyrics, vocals, and instrumental tracks using Tencent AI Lab's SongGeneration (LeVo) model. [NVIDIA ONLY]
6Morpheus6/SongGeneration-Studiov3.7updated 1mo ago
AgentsViewFeatured
Browse, search, and analyze sessions from Claude Code, Codex, and other coding agents locally.
cocktailpeanut/agentsview.pinokiov8.0updated 1mo ago
GeoLibreFeatured
A free, open-source GIS for visualizing, exploring, and analyzing geospatial data locally in your browser.
cocktailpeanut/geolibre.pinokiov7.0updated 27d ago
Uncensored Local StudioFeatured
Run image generation, GGUF language models, Whisper speech recognition, and Kokoro speech synthesis locally from one offline studio.
cocktailpeanut/uncensored-local-studio.pinokiov8.0updated 20d ago
MuScriptorFeatured
Local multi-instrument audio-to-MIDI transcription from Kyutai and Mirelo.
cocktailpeanut/muscriptor.pinokiov8.0.0updated 1mo ago
MaestroFeatured
An all-in-one, 100% local AI creative studio, director, and multi-track editor. Generate with MiniMax H3, LTX-2.5/2.3, Wan, Flux, Qwen, and more; turn an idea or song into a planned production; then finish it on the timeline. Requires an NVIDIA GPU (6GB+ VRAM).
Blizaine/Maestrov8.0updated 27 min ago
UnderfitFeatured
LoRA fine-tuning dashboard for Stable Audio 3
cocktailpeanut/underfit.pinokiov7.0updated 1mo ago


