Today's Image & Video Generation: Fastest-Growing Projects — August 18, 2026
Today's the Image & Video Generation space on GitHub, we see a strong emphasis on creative and artistic uses of AI, with several repositories gaining traction for their unique approaches to image and video generation. Notably, developers are leveraging Codex to generate visually appealing illustrations from everyday photos and enhancing portrait photography through advanced lighting techniques.
*yokel1121/muyang-illustration-skills* offers 25 fixed illustration styles for Codex with prompt recipes and visual references, making it easier for users to create specific types of illustrations. With a Growth Score of 20.57 and 183 stars, this repository is growing due to its comprehensive set of tools that cater to artists and designers looking to enhance their creative process.
*dacnay816y62-hub/photo-revival* transforms ordinary photos into poetic, hand-drawn illustrations with a white-paper style. This project has garnered 435 stars and a Growth Score of 17.91, likely due to its unique approach that appeals to those interested in artistic photo manipulation.
*Sateezg/codex-bridge* provides image generation capabilities through the Codex CLI without requiring an OpenAI API key, integrating GPT-based subagents like gpt-image-2 and a GPT-5 agent. With 17.17 Growth Score and 255 stars, it stands out for its user-friendly interface and broad compatibility with existing tools.
*shixinnt/codex-image-context-runtime* is a Codex plugin that supports context-bounded image generation and inspection through a local MCP runtime. Its strong growth score of 14.86 and 100 stars suggest that it meets a need for more controlled, contextual image generation environments among developers.
*oil-oil/dsh-vision* aims to provide near-native understanding of images within the DeepSeek Harness framework. With a Growth Score of 12.00 and 63 stars, this repository is growing as users seek advanced image analysis tools that can integrate seamlessly into existing workflows.
*moskoo/xxg-portrait-rebuild-light* focuses on reconstructing portrait photos with enhanced lighting effects and clean skin textures using director-led techniques. Its Growth Score of 8.79 and 51 stars indicate its appeal to photographers and photo editors looking for sophisticated image editing tools.
*Tencent-Hunyuan/Hunyuan3D-Buffalo1.0* presents a unified multimodal model for scalable 3D generation, understanding, and editing. With 204 stars and a Growth Score of 7.58, it captures the interest of developers working on complex 3D projects requiring advanced AI capabilities.
*mtsee/unicut* is an open-source video editing tool powered by AI that streamlines video creation processes. Its steady growth with a score of 7.34 and 184 stars reflects its utility for content creators seeking efficient video production solutions.
*24029100313/RA-Bench* offers a benchmark designed to detect AI-generated videos in real-world crisis settings, anchoring detection in real events. With 6.10 Growth Score and 37 stars, it addresses the growing need for reliable methods to verify video authenticity.
Finally, *BeatAPI/awesome-minimax-h3-prompts* provides curated prompts for generating various types of videos with MiniMax H3, including cinematic clips, ads, anime, UGC, product videos, and more. Its 6.03 Growth Score and 109 stars highlight its value to creators seeking inspiration or standardized templates for video generation.
These repositories showcase a diverse range of applications in image and video generation, from artistic illustration to advanced 3D modeling and real-world crisis detection, reflecting the broad impact of AI technologies across different domains.
*yokel1121/muyang-illustration-skills* offers 25 fixed illustration styles for Codex with prompt recipes and visual references, making it easier for users to create specific types of illustrations. With a Growth Score of 20.57 and 183 stars, this repository is growing due to its comprehensive set of tools that cater to artists and designers looking to enhance their creative process.
*dacnay816y62-hub/photo-revival* transforms ordinary photos into poetic, hand-drawn illustrations with a white-paper style. This project has garnered 435 stars and a Growth Score of 17.91, likely due to its unique approach that appeals to those interested in artistic photo manipulation.
*Sateezg/codex-bridge* provides image generation capabilities through the Codex CLI without requiring an OpenAI API key, integrating GPT-based subagents like gpt-image-2 and a GPT-5 agent. With 17.17 Growth Score and 255 stars, it stands out for its user-friendly interface and broad compatibility with existing tools.
*shixinnt/codex-image-context-runtime* is a Codex plugin that supports context-bounded image generation and inspection through a local MCP runtime. Its strong growth score of 14.86 and 100 stars suggest that it meets a need for more controlled, contextual image generation environments among developers.
*oil-oil/dsh-vision* aims to provide near-native understanding of images within the DeepSeek Harness framework. With a Growth Score of 12.00 and 63 stars, this repository is growing as users seek advanced image analysis tools that can integrate seamlessly into existing workflows.
*moskoo/xxg-portrait-rebuild-light* focuses on reconstructing portrait photos with enhanced lighting effects and clean skin textures using director-led techniques. Its Growth Score of 8.79 and 51 stars indicate its appeal to photographers and photo editors looking for sophisticated image editing tools.
*Tencent-Hunyuan/Hunyuan3D-Buffalo1.0* presents a unified multimodal model for scalable 3D generation, understanding, and editing. With 204 stars and a Growth Score of 7.58, it captures the interest of developers working on complex 3D projects requiring advanced AI capabilities.
*mtsee/unicut* is an open-source video editing tool powered by AI that streamlines video creation processes. Its steady growth with a score of 7.34 and 184 stars reflects its utility for content creators seeking efficient video production solutions.
*24029100313/RA-Bench* offers a benchmark designed to detect AI-generated videos in real-world crisis settings, anchoring detection in real events. With 6.10 Growth Score and 37 stars, it addresses the growing need for reliable methods to verify video authenticity.
Finally, *BeatAPI/awesome-minimax-h3-prompts* provides curated prompts for generating various types of videos with MiniMax H3, including cinematic clips, ads, anime, UGC, product videos, and more. Its 6.03 Growth Score and 109 stars highlight its value to creators seeking inspiration or standardized templates for video generation.
These repositories showcase a diverse range of applications in image and video generation, from artistic illustration to advanced 3D modeling and real-world crisis detection, reflecting the broad impact of AI technologies across different domains.