R_REDDYX.XYZ
MultimodalTypeScript★ 233⑂ 49 forksMIT

workingprototype/Fuobe

OPEN CREATIVE STUDIO ON COMFYUI

Local studio for generating images, video, 3D and audio based on ComfyUI. Runs on open weights on your GPU or in the cloud via API.

// KEY FEATURES

  • 108 camera motions, 255 effects and 132 apps in grouped menus
  • 3D generation from photo, text or object + Body Double with motion capture
  • Assist agent plans task in chat, shows cost and runs on approval
  • Layers for editing, Kinetic Motion for motion graphics and marketing templates

// DEEP DIVE: HOW IT WORKS

Fuobe is a web interface on top of ComfyUI, combining 29 studios for working with images, video, 3D and audio. The user chooses mode: local run of open-weight models on their GPU (tested on RTX 5090) or cloud via fal.ai and Replicate keys. All previews and renders execute inside ComfyUI, while control runs through a browser dashboard with the Assist chat agent.

Inside — 108 camera motion presets (any two can be mixed), 255 effects (weather, transitions, lenses, grading), 132 one-click apps (outfit swap, relighting, upscale, restore) and 62 Fuobe FX tools for VFX pipeline: plate recast, mattes, retime, 3D extension. Added 5 Image-to-3D modes (photo, multiview, prompt, object, splat), Body Double for 3D body with motion capture, Kinetic Motion for branded typography, layers for editing flat images, and 30 film looks for MOSS-Image.

Requires CUDA GPU for local mode or fal.ai/Replicate API keys for cloud. Interface in TypeScript, deploys on top of existing ComfyUI installation. No built-in models — weights download on demand. Assist agent estimates run cost before start, but does not guarantee accuracy for cloud providers.

// USE CASES

Best for: For artists, videomakers and FX teams for local production on ComfyUI

  • Generate ad video with camera motion and effects from single product photo
  • Create animated 3D character model from photo and apply motion capture from video
  • Give agent task 'make UGC video for TikTok' and get finished render with cost estimate
#ai-video#comfyui#creative-tools#generative-ai#ideogram#image-generation#image-to-3d#image-to-video#ltx-video#minimax-h3#nextjs#open-source
Open on GitHub →

New repositories every 30 minutes

REDDYX AI scans GitHub 24/7 and ships the best AI/ML/Web3 projects to Telegram.

Join on Telegram
← Full catalog·Full index

// SIMILAR REPOSITORIES

MULTIMODAL DECISION MODEL★ 918REMOVING AI WATERMARKS AND METADATA★ 6403LTX-2: DIT MODEL FOR SYNCHRONIZED AUDIO AND VIDEO★ 10312EMBEDDED VECTOR DB FOR MULTIMODAL AI★ 11947OPEN SOURCE AI WORKSPACE FOR CREATORS★ 14387PROFESSIONAL AI ENGINE FOR VISUAL MEDIA★ 29106

← FULL CATALOG

workingprototype/Fuobe — Open Creative Studio ON Comfyui