Scenario ↗
Multimedia creation & assistantCreate and transform images, videos, audio, 3D models, materials and skyboxes. The studio shows the models available to your account and the settings for each generation.
Imagine. Generate. Create.
From images to sound, video to 3D. Generate and refine your creations in one project.
Home · the starting point
Keep images, sounds, scenes and references in one project. Switch workshops, arrange your panels and keep your files in open formats.
Image · creation and editing
Generate an image. Select what matters. Refine it, layer by layer.
3D · composition and animation
Compose a scene. Animate a character. Put your ideas in motion.
Workshop · .mp4 output
Two monitors, a multitrack timeline, and models that produce the shots.
Workshop · .wav output
Compose and edit sound across several tracks, in the same window as the rest.
Materials · channels and 3D preview
A material, not an image: each channel is produced then checked separately.
Skies · environment and light
A sky that serves as a backdrop and, at the same time, as a light source.
AI models · local & cloud
Create an image, animate a character, dictate an idea or ask for help. Choose the right model for each task, on your computer or remotely.
In the cloud
Your accounts, keys and models. Generation and the assistant each keep their own model choice.
Create and transform images, videos, audio, 3D models, materials and skyboxes. The studio shows the models available to your account and the settings for each generation.
Turn text, an image or multiple views into 3D. Texture, refine, stylize, segment or simplify a mesh; prepare its rig and animations. Generate and edit the source images too.
Conversation, creative help and scripting. Image analysis depends on the selected model.
The cloud catalogue changes with services and account permissions. Each provider’s credits and terms apply.
On your computer
Every use, every model: the complete local catalogue.
Chat, understand a request and prepare actions in the studio.
Write and revise scripts for your scenes.
Generation, editing, inpainting, outpainting, references and guidance. Compatible models also support materials.
Isolate a region from a point or rectangle, then refine its mask.
Create from text, animate an image or transform a video, depending on the model.
Create or transform music; generate sound to accompany a video.
Turn text or an image into a 3D model, then modify its geometry.
Create a panoramic environment from text or an image.
Automatically prepare a character’s skeleton and skinning weights.
Retrieve the meaning of your conversations and stored information.
Generate motion from text. Listed in the catalogue; download blocked, user import required.
Turn speech into text without sending audio to the cloud.
Your library too
Find your compatible Ollama models or import a GGUF file for the assistant. You control which model is used.
Local models require installation. Availability depends on the engine, memory and hardware; each model has its own licence. Kimodo is listed but its download is blocked. AI upscaling and vectorization currently have no declared engine in this catalogue.
Your models, your project
Choose models for what you want to create. Local or cloud, the tools stay together in one studio.
Arrange panels and tabs, or detach them into another window. Each workshop remembers its layout for the next session.
Documents rebuild from their saved state, including in detached panels. The same foundation supports saving and undo.
Browse files even without the studio. Manage versions, branches and conflicts with Git. Import, export and the library find the same media, even after moves outside the app.
Find each model’s options in one panel, for every kind of generation.
Follow your generations and find their results alongside other media in your project.
The system encrypts your keys. The interface receives connection status without receiving secrets.
Download and manage compatible models. Availability depends on the operation, model and your computer’s resources.
Documents, images, scenes, animation, settings and files: connect an external assistant to the studio’s published MCP actions.
Thumbnails, waveforms, heavy geometry and video encoding use the GPU or tasks separate from the interface.
Documentation & demonstration
Installation
Download the published version for your computer.
Apple Silicon · .dmg
Release to comeIntel · .dmg
Release to comex64 · .exe
Release to comex86_64 · .AppImage
Release to comex86_64 · .deb
Release to comeThe installer is governed by the terms of use; the source code, by the PolyForm Noncommercial 1.0.0 licence. Cloud services use your account and credits. Local models retain their own terms of use.
git clone https://github.com/pasquelin/AIDesktopStudio.git
pnpm install
pnpm start