Today's Image & Video Generation: Fastest-Growing Projects — May 24, 2026
Today's radar in the Image & Video Generation space highlights a diverse range of projects catering to various user needs, from professional designers looking for high-quality AI-generated media to enthusiasts seeking accessible and customizable tools. Among these, ByteDance’s Lance stands out with its comprehensive approach to multimodal tasks, while other projects like Stable-Diffusion-WebUI offer more specialized features tailored towards specific use cases such as image generation via web interfaces.
bytedance/Lance is a lightweight native unified multimodal model designed for understanding, generating, and editing images and videos. With a growth score of 66.60, it appears to be rapidly gaining traction among developers interested in leveraging advanced AI capabilities without significant computational overhead. This tool's versatility and its ability to handle multiple tasks within a single framework make it an attractive option for those looking to streamline their multimedia processing workflows.
BasZ4ll/Stable-Diffusion-WebUI is a web interface for Stable Diffusion, providing users with guides on how to download free versions of SD WebUI Forge launcher and optimize settings like low VRAM usage and xformers command line arguments. Despite having fewer commits in the last 30 days, its growth score of 30.29 indicates steady interest from developers who value comprehensive documentation and ease-of-use features for Stable Diffusion models. The repository's extensive star count (400) suggests a strong community engagement.
mrslimslim/gpt-image-canvas offers a local professional AI canvas built with tldraw, enabling users to create high-quality images using advanced AI capabilities. With 524 stars and a growth score of 24.10, this project is witnessing significant interest from designers and developers who appreciate the tool’s integration of AI-driven features within an intuitive interface designed for ease of use. The continuous development activity (84 commits in the last month) reflects ongoing improvements to enhance user experience.
helloianneo/ian-handdrawn-ppt generates full-page images, including 21:9 cover pages and 16:9 content illustrations, ideal for creating visually appealing presentations with a hand-drawn touch. With 880 stars and a growth score of 19.62, this project resonates particularly well among Chinese-speaking users who value aesthetically pleasing presentation materials that can be exported as PNG files. The single commit in the last month might indicate stabilization after rapid development phases.
RareSense/Nova3D enables editable, part-aware 3D generation from text or reference images through an open-source client for nova3d.xyz. With a growth score of 15.58 and 294 stars, it attracts developers seeking to leverage AI in creating detailed 3D models with precise control over individual components. The consistent development activity (26 commits in the last month) suggests active refinement and expansion of its feature set.
ailuntx/Thinking-with-Visual-Primitives is an archived snapshot project that was likely part of a larger initiative focusing on visual primitive-based AI thinking. With 142 stars but a low growth score of 6.58, it seems to have reached a stable phase with limited recent development activity, indicating its primary value lies in historical context or as a reference for similar projects.
Hao0321/ai-media-generator automates the creation of high-quality AI-generated media across various platforms using Claude Code Skill prompts. Featuring 31 stars and a growth score of 5.64, it appeals to users aiming to craft professional-grade image, video, and music prompts effortlessly. The frequent commits (12 in the last month) suggest ongoing enhancements to support more platforms and improve user experience.
jieyefriic/nbcraft provides a multi-backend CLI for generating images and videos using Gemini, DashScope, Volcengine Ark, and OpenAI’s gpt-image-2 models. With 3.62 growth score and 153 stars, this tool caters to developers preferring command-line interfaces and seeking flexibility across different AI service providers. The minimal development activity (only two commits in the last month) may indicate that nbcraft has reached a stable state post-development.
gpt-img-2/ai-image-prompt-cookbook compiles Chinese-language AI image prompt libraries for diverse applications such as fashion, e-commerce product images, and social media posts. With 26 stars and a growth score of 3.61, it targets a niche but engaged community interested in localized content creation through AI. The recent updates (three commits) reflect continuous efforts to expand the library’s scope.
microsoft/Lens is a text-to-image diffusion model with a parameter count of 3.8B designed for high-quality image generation, outperforming models like FLUX and SD3 while requiring less training compute. With just 60 stars and a growth score of 2.34,
bytedance/Lance is a lightweight native unified multimodal model designed for understanding, generating, and editing images and videos. With a growth score of 66.60, it appears to be rapidly gaining traction among developers interested in leveraging advanced AI capabilities without significant computational overhead. This tool's versatility and its ability to handle multiple tasks within a single framework make it an attractive option for those looking to streamline their multimedia processing workflows.
BasZ4ll/Stable-Diffusion-WebUI is a web interface for Stable Diffusion, providing users with guides on how to download free versions of SD WebUI Forge launcher and optimize settings like low VRAM usage and xformers command line arguments. Despite having fewer commits in the last 30 days, its growth score of 30.29 indicates steady interest from developers who value comprehensive documentation and ease-of-use features for Stable Diffusion models. The repository's extensive star count (400) suggests a strong community engagement.
mrslimslim/gpt-image-canvas offers a local professional AI canvas built with tldraw, enabling users to create high-quality images using advanced AI capabilities. With 524 stars and a growth score of 24.10, this project is witnessing significant interest from designers and developers who appreciate the tool’s integration of AI-driven features within an intuitive interface designed for ease of use. The continuous development activity (84 commits in the last month) reflects ongoing improvements to enhance user experience.
helloianneo/ian-handdrawn-ppt generates full-page images, including 21:9 cover pages and 16:9 content illustrations, ideal for creating visually appealing presentations with a hand-drawn touch. With 880 stars and a growth score of 19.62, this project resonates particularly well among Chinese-speaking users who value aesthetically pleasing presentation materials that can be exported as PNG files. The single commit in the last month might indicate stabilization after rapid development phases.
RareSense/Nova3D enables editable, part-aware 3D generation from text or reference images through an open-source client for nova3d.xyz. With a growth score of 15.58 and 294 stars, it attracts developers seeking to leverage AI in creating detailed 3D models with precise control over individual components. The consistent development activity (26 commits in the last month) suggests active refinement and expansion of its feature set.
ailuntx/Thinking-with-Visual-Primitives is an archived snapshot project that was likely part of a larger initiative focusing on visual primitive-based AI thinking. With 142 stars but a low growth score of 6.58, it seems to have reached a stable phase with limited recent development activity, indicating its primary value lies in historical context or as a reference for similar projects.
Hao0321/ai-media-generator automates the creation of high-quality AI-generated media across various platforms using Claude Code Skill prompts. Featuring 31 stars and a growth score of 5.64, it appeals to users aiming to craft professional-grade image, video, and music prompts effortlessly. The frequent commits (12 in the last month) suggest ongoing enhancements to support more platforms and improve user experience.
jieyefriic/nbcraft provides a multi-backend CLI for generating images and videos using Gemini, DashScope, Volcengine Ark, and OpenAI’s gpt-image-2 models. With 3.62 growth score and 153 stars, this tool caters to developers preferring command-line interfaces and seeking flexibility across different AI service providers. The minimal development activity (only two commits in the last month) may indicate that nbcraft has reached a stable state post-development.
gpt-img-2/ai-image-prompt-cookbook compiles Chinese-language AI image prompt libraries for diverse applications such as fashion, e-commerce product images, and social media posts. With 26 stars and a growth score of 3.61, it targets a niche but engaged community interested in localized content creation through AI. The recent updates (three commits) reflect continuous efforts to expand the library’s scope.
microsoft/Lens is a text-to-image diffusion model with a parameter count of 3.8B designed for high-quality image generation, outperforming models like FLUX and SD3 while requiring less training compute. With just 60 stars and a growth score of 2.34,