Today's Image & Video Generation: Fastest-Growing Projects — April 25, 2026
Today's Image & Video Generation, we saw a surge in tools that leverage AI to generate high-quality images and videos. The trend is shifting towards more accessible and user-friendly interfaces, with several projects offering web UIs or CLI tools for easy integration. Meanwhile, research-focused repositories continue to push the boundaries of multimodal understanding and text-to-image generation.
Lidge-jun's ima2-gen has taken the top spot with a growth score of 89.00 and 67 stars, thanks to its minimal yet powerful CLI and web UI for OpenAI GPT Image 2 generation. Its dual authentication system, supporting both API keys and OAuth via ChatGPT, makes it an attractive option for developers.
JD-opensource's JoyAI-Image boasts an impressive 1,951 stars and a growth score of 51.68, driven by its unified multimodal foundation model for image understanding, text-to-image generation, and instruction-guided image editing. Its versatility and robust capabilities have made it a go-to choice for researchers and developers alike.
Snapotter-hq's SnapOtter has garnered significant attention with a growth score of 29.26 and 887 stars, offering a self-hosted image manipulator with over 45 tools, local AI, and pipelines in a single Docker container. Its emphasis on data privacy and security has resonated with users seeking control over their images.
Baidu's ERNIE-Image, with a growth score of 26.32 and 375 stars, presents an open text-to-image generation model developed by the ERNIE-Image team at Baidu. Its single-stream Diffusion Transformer (DiT) architecture has achieved state-of-the-art performance among open-weight text-to-image models.
Kangarooking's design-image-studio has seen modest growth with a score of 14.86 and 88 stars, focusing on directly generating high-quality design images from vague visual demands. Its unique approach to image generation has piqued the interest of designers and developers.
OpenMOSS's MOSS-VL boasts a growth score of 14.56 and 231 stars, driven by its core multimodal model series dedicated to visual understanding within the OpenMOSS ecosystem. Researchers have taken notice of its potential in advancing visual comprehension.
ShandaAI's AlayaRenderer has secured a growth score of 12.72 and 528 stars with its AI-native Renderer for Games and Virtual Worlds. Its innovative approach to rendering has captivated the attention of game developers and researchers exploring virtual worlds.
Gulucaptain's Camera-Transformer-1, with a growth score of 11.57 and 167 stars, offers a unique camera control system that understands user intent and generates videos with precise spatial awareness. Its application in video production has sparked interest among creators.
Jiangchaokang's awesome-generative-models has grown steadily with a score of 7.46 and 100 stars, providing weekly updates on the latest diffusion generation models across various fields. Researchers have come to rely on its curated list for staying current with the rapidly evolving landscape of generative models.
NVlabs' PixelDiT rounds out our list with a growth score of 6.06 and 252 stars, showcasing its Pixel Diffusion Transformers for Image Generation. Although it has seen limited activity recently, its potential in advancing image generation techniques remains promising.
Lidge-jun's ima2-gen has taken the top spot with a growth score of 89.00 and 67 stars, thanks to its minimal yet powerful CLI and web UI for OpenAI GPT Image 2 generation. Its dual authentication system, supporting both API keys and OAuth via ChatGPT, makes it an attractive option for developers.
JD-opensource's JoyAI-Image boasts an impressive 1,951 stars and a growth score of 51.68, driven by its unified multimodal foundation model for image understanding, text-to-image generation, and instruction-guided image editing. Its versatility and robust capabilities have made it a go-to choice for researchers and developers alike.
Snapotter-hq's SnapOtter has garnered significant attention with a growth score of 29.26 and 887 stars, offering a self-hosted image manipulator with over 45 tools, local AI, and pipelines in a single Docker container. Its emphasis on data privacy and security has resonated with users seeking control over their images.
Baidu's ERNIE-Image, with a growth score of 26.32 and 375 stars, presents an open text-to-image generation model developed by the ERNIE-Image team at Baidu. Its single-stream Diffusion Transformer (DiT) architecture has achieved state-of-the-art performance among open-weight text-to-image models.
Kangarooking's design-image-studio has seen modest growth with a score of 14.86 and 88 stars, focusing on directly generating high-quality design images from vague visual demands. Its unique approach to image generation has piqued the interest of designers and developers.
OpenMOSS's MOSS-VL boasts a growth score of 14.56 and 231 stars, driven by its core multimodal model series dedicated to visual understanding within the OpenMOSS ecosystem. Researchers have taken notice of its potential in advancing visual comprehension.
ShandaAI's AlayaRenderer has secured a growth score of 12.72 and 528 stars with its AI-native Renderer for Games and Virtual Worlds. Its innovative approach to rendering has captivated the attention of game developers and researchers exploring virtual worlds.
Gulucaptain's Camera-Transformer-1, with a growth score of 11.57 and 167 stars, offers a unique camera control system that understands user intent and generates videos with precise spatial awareness. Its application in video production has sparked interest among creators.
Jiangchaokang's awesome-generative-models has grown steadily with a score of 7.46 and 100 stars, providing weekly updates on the latest diffusion generation models across various fields. Researchers have come to rely on its curated list for staying current with the rapidly evolving landscape of generative models.
NVlabs' PixelDiT rounds out our list with a growth score of 6.06 and 252 stars, showcasing its Pixel Diffusion Transformers for Image Generation. Although it has seen limited activity recently, its potential in advancing image generation techniques remains promising.