Launching today
Contsy Studio
Consistent AI character photos and talking videos, locally
2 followers
Consistent AI character photos and talking videos, locally
2 followers
Create a consistent AI character once, then get endless 1080p photos and talking videos, 100% on your own PC. One-click Windows install, built on ComfyUI.
Free
Launch Team
AWS Startup AdvisorPersonalized AWS guidance that meets you where you’re already building
Promoted
Hi Product Hunt! I'm Karol, the solo founder of Contsy and creator of Contsy Studio.
I started building this in mid-October while attending technical secondary school. I use AI coding tools to develop the application, bringing ComfyUI and existing AI models into one interface.
Contsy Studio lets you train a recurring character, create photos in different scenes, and generate talking videos with speech and lip sync. AI Series plans batches from a prompt, and Autopilot prepares content in the background for review. It does not publish to social platforms.
The local application is free and open source. It runs on Windows 11 with an NVIDIA GPU. The Lite profile lists 8 GB VRAM, 32 GB RAM and about 130 GB of SSD space. Model downloads are large, and video generation takes time. Underlying models retain their own licenses.
I'm looking for external testers. Is the installation guide clear, and how well does character consistency work for your use case? If you try it, please include your GPU and VRAM with feedback.
Source and installation:
https://github.com/karol-kosciolek/contsy-studio
Thanks for taking a look!