🎬 MiniMax-H3 Studio
Reference-to-Video & Text-to-Video powered by
MiniMaxAI/MiniMax-H3
via Hugging Face Inference Providers. No GPU required.
2 10
Resolution
Inference provider
💡 Tip: a clear subject + simple motion + explicit camera direction (dolly / pan / orbit) tends to give the cleanest results.
2 10
Resolution
Inference provider
💡 Tip: cinematic keywords (cinematic, 35mm film, shallow depth of field, volumetric lighting) help the model lock onto a look.
Click any example to load it into the T2V tab.
Built with Gradio on Hugging Face Spaces · Inference via
HF Inference Providers ·
Model: MiniMaxAI/MiniMax-H3