Testing MiniMax I2V: ComfyUI Workflow and Booru Tag Tips
Infamous_Campaign687 · reddit · 2026-08-04
The author shares a hands-on test of the default MiniMax Image-to-Video (I2V) workflow within ComfyUI, demonstrating a specific video generation pipeline.
- Workflow: Involves recording a few seconds of screen capture, extracting the last frame, running I2V generation with an LLM-enhanced prompt, and finally stitching the recording and generated video together using ffmpeg.
- Tips: Warns users to be careful when using Booru tags (an anime image tagging system) with the model to avoid unexpected generation results.
Related event: MiniMax Image-to-Video Model Tested(2 posts)→
More from Multimodal
- MiniMax Video Model's True Strength Lies in Long-Form Storytelling — cocktailpeanut · 2026-08-06
- H3 Model Generates Blocky and Distorted Images, Users Debug Node Settings — FastIce8391 · 2026-08-06
- Minimax video model runs locally: 4080 generates 1280x720 in ~30 mins — cocktailpeanut · 2026-08-06
- FLUX3's edge: 20 seconds generation time — cocktailpeanut · 2026-08-06
- H3 Video Model Review: Great Motion Control but Slow, 15s Takes 800s on A100 — ReporterRemote6713 · 2026-08-06
- ComfyUI Oasis v1.6: Seamless Long Video Chaining via Motion Context — Sad_Berry_4621 · 2026-08-06