Today's Image & Video Generation: Fastest-Growing Projects — October 23, 2026
Today's the Image & Video Generation space on GitHub, we see a diverse range of projects gaining traction from developers and researchers alike. Notably, there's significant interest in open-source solutions that enhance image quality and streamline creative processes for artists. The top tools this week include a pixel art editor and an AI-driven image super-resolution tool.
yangsir2077/pixel-loom is an open-source pixel art editing software that allows users to create and edit animations using AI technology through the MCP framework, which supports Windows portable editions. With its steady growth score of 15.69 and a modest but growing number of stars at 80, this tool stands out for its unique approach to integrating AI into traditional pixel art creation.
isixe/bg2x is described as a privacy-first AI image super-resolution (upscaling) tool that enhances the resolution of images without compromising user data. It has garnered significant attention with 14.06 growth points and an impressive 119 stars, likely due to its commitment to user privacy while delivering high-quality image enhancement features.
speridlabs/iris-3b is a diffusion transformer model that generates pixel-space images directly without the need for a Variational Autoencoder (VAE). This project also includes fine-tuned models capable of monocular depth estimation and 4x image restoration, showcasing its versatility in computer vision tasks. With a growth score of 9.22 and 200 stars, Iris-3B's comprehensive approach to image generation and restoration is clearly resonating with the developer community.
SimonGiebenhain/video-diffusion-avatars provides official code for transforming video diffusion models into controllable 3D avatars, offering a unique solution in the realm of avatar creation from video data. Its growth score of 6.21 and 51 stars indicate its relevance to developers interested in creating personalized or interactive digital characters based on video inputs.
harshpreet931/harshpreet-hand details an experiment where Claude AI is trained with handwritten college notes to create a font that closely resembles the original handwriting, demonstrating the potential of generative models in personalizing text and design. With 47 stars and a growth score of 4.67, this project highlights the creative applications of AI in transforming physical handwriting into digital fonts.
PolyU-VCLab/HarnessIR is designed to harness multimodal foundation models for universal real-world image restoration tasks, aiming to improve image quality across various scenarios. Its growth score of 3.78 and 34 stars suggest that developers are interested in its broad applicability and potential for enhancing image clarity and detail.
ejsyk3xdp3t2o/wan3.0video-wan-3.0-video-api offers an API gateway providing access to models capable of generating video content at different resolutions, aligning with the growing demand for scalable solutions in video generation. With a growth score of 3.16 and 93 stars, this project is seeing traction due to its compatibility with OpenAI standards and ease of integration.
VectorSpaceLab/InfoVis presents an innovative model that can generate and edit slide images alongside a benchmarking suite for evaluating such models. Its growth score of 3.14 and 40 stars reflect the interest in advanced tools that streamline visual communication, particularly in academic or professional presentations.
worldbench/Orbia explores whether generated videos form a consistent three-dimensional world, addressing foundational questions about AI-generated content's coherence and realism. With a growth score of 2.80 and 48 stars, this project is attracting attention for its theoretical contributions to the field of video generation.
sunshy-1/StyleForge focuses on eliciting inherent style transfer capabilities in step-distilled diffusion models, offering insights into fine-tuning these models for diverse creative applications. Its growth score of 2.21 and 24 stars indicate a niche but growing interest among developers exploring the boundaries of AI-driven artistic expression.
These projects collectively highlight the expanding landscape of image and video generation technologies, each contributing unique value to the developer ecosystem through innovative approaches and practical solutions.
yangsir2077/pixel-loom is an open-source pixel art editing software that allows users to create and edit animations using AI technology through the MCP framework, which supports Windows portable editions. With its steady growth score of 15.69 and a modest but growing number of stars at 80, this tool stands out for its unique approach to integrating AI into traditional pixel art creation.
isixe/bg2x is described as a privacy-first AI image super-resolution (upscaling) tool that enhances the resolution of images without compromising user data. It has garnered significant attention with 14.06 growth points and an impressive 119 stars, likely due to its commitment to user privacy while delivering high-quality image enhancement features.
speridlabs/iris-3b is a diffusion transformer model that generates pixel-space images directly without the need for a Variational Autoencoder (VAE). This project also includes fine-tuned models capable of monocular depth estimation and 4x image restoration, showcasing its versatility in computer vision tasks. With a growth score of 9.22 and 200 stars, Iris-3B's comprehensive approach to image generation and restoration is clearly resonating with the developer community.
SimonGiebenhain/video-diffusion-avatars provides official code for transforming video diffusion models into controllable 3D avatars, offering a unique solution in the realm of avatar creation from video data. Its growth score of 6.21 and 51 stars indicate its relevance to developers interested in creating personalized or interactive digital characters based on video inputs.
harshpreet931/harshpreet-hand details an experiment where Claude AI is trained with handwritten college notes to create a font that closely resembles the original handwriting, demonstrating the potential of generative models in personalizing text and design. With 47 stars and a growth score of 4.67, this project highlights the creative applications of AI in transforming physical handwriting into digital fonts.
PolyU-VCLab/HarnessIR is designed to harness multimodal foundation models for universal real-world image restoration tasks, aiming to improve image quality across various scenarios. Its growth score of 3.78 and 34 stars suggest that developers are interested in its broad applicability and potential for enhancing image clarity and detail.
ejsyk3xdp3t2o/wan3.0video-wan-3.0-video-api offers an API gateway providing access to models capable of generating video content at different resolutions, aligning with the growing demand for scalable solutions in video generation. With a growth score of 3.16 and 93 stars, this project is seeing traction due to its compatibility with OpenAI standards and ease of integration.
VectorSpaceLab/InfoVis presents an innovative model that can generate and edit slide images alongside a benchmarking suite for evaluating such models. Its growth score of 3.14 and 40 stars reflect the interest in advanced tools that streamline visual communication, particularly in academic or professional presentations.
worldbench/Orbia explores whether generated videos form a consistent three-dimensional world, addressing foundational questions about AI-generated content's coherence and realism. With a growth score of 2.80 and 48 stars, this project is attracting attention for its theoretical contributions to the field of video generation.
sunshy-1/StyleForge focuses on eliciting inherent style transfer capabilities in step-distilled diffusion models, offering insights into fine-tuning these models for diverse creative applications. Its growth score of 2.21 and 24 stars indicate a niche but growing interest among developers exploring the boundaries of AI-driven artistic expression.
These projects collectively highlight the expanding landscape of image and video generation technologies, each contributing unique value to the developer ecosystem through innovative approaches and practical solutions.