Innovate Futures @ Benji
patreon
Flux Kontext + AniSora 2.0 AI Video - Create the Anime Series You Dreamed of as a Kid
🕑 Added 2025-07-16 13:00:17 +0000 UTC
Tutorial Video: https://youtu.be/yMAcyQLLsQI
In this video, we explore the new AniSora AI video model , a fine-tuned version based on WAN 2.1 and developed by Bilibili (also known as B Station), one of China’s largest platforms for anime, manga, and user-generated content. This open-source model is specifically trained to generate cartoon-style and anime videos , making it a powerful tool for creators looking to produce animated content using AI. The video walks through how to set up and use AniSora in ComfyUI , including downloading and installing the model, integrating it with tools like Flux Context for image-to-anime conversion, and generating short clips using text prompts and reference images. The creator also shares real-world testing results, performance benchmarks, and honest feedback about the model’s strengths—like its fast generation speed at lower resolutions—and limitations, such as occasional artifacts and inconsistencies in motion and detail.
Who is This Content Suitable For?
This content is ideal for:
Anime and cartoon creators interested in leveraging AI for video production.
AI developers and researchers working with video diffusion models , especially those focused on animated content .
ComfyUI users looking to expand their toolkit with new AI video models like AniSora and WAN 2.1 .
Content creators who want to experiment with image-to-video generation , style transfer , or text-to-anime workflows .
Anyone curious about open-source AI models from Chinese tech platforms like Bilibili.
Why Does This Matter?
The rise of AI-generated animation opens new doors for independent creators, allowing them to produce high-quality animated content without needing large teams or expensive software. AniSora , being an open-source model trained specifically on anime and cartoon data, offers a unique opportunity for creators to generate stylized videos quickly and efficiently. Understanding how to integrate and optimize this model within tools like ComfyUI helps creators streamline their workflow, test creative ideas rapidly, and stay ahead in the evolving world of generative AI for animation.
Anisora Github
https://github.com/bilibili/Index-anisora
For ComfyUI
--------------------------------
Model:
Safetensors:
Wan2_1-Anisora-I2V-480P-14B_fp16.safetensors
https://huggingface.co/Kijai/WanVideo_comfy/blob/main/Wan2_1-Anisora-I2V-480P-14B_fp16.safetensors
Wan2_1-Anisora-I2V-480P-14B_fp8_e4m3fn.safetensors
GGUF:
hanzogak/Wan2.1-Anisora-14B-I2V-480P-GGUF
https://huggingface.co/hanzogak/Wan2.1-Anisora-14B-I2V-480P-GGUF/tree/main
Wan2_1-Anisora-I2V-480P-14B_Q8_0.gguf
https://huggingface.co/Kijai/WanVideo_comfy/blob/main/Wan2_1-Anisora-I2V-480P-14B_Q8_0.gguf
vae
Clip_vision_h
https://huggingface.co/Comfy-Org/Wan_2.1_ComfyUI_repackaged/tree/main/split_files/clip_vision
text encoder
https://huggingface.co/Kijai/WanVideo_comfy/blob/main/umt5-xxl-enc-fp8_e4m3fn.safetensors
Example Workflows in this tutorial attached below
