I will build a custom comfyui ai video workflow with minimax h3 wan and ltx

MiniMax H3 through open weights or Comfy Partner Nodes, plus Wan and LTX. Seedance or Kling can be connected when you provide valid API access. I confirm the best model after reviewing your assets, hardware, access, and output goal.

H3 supports text-to-video and reference-to-video with native audio options, 5-15 second clips, and targets up to 2K. Exact duration, resolution, audio, speed, and quality depend on the selected deployment and are validated for your agreed package.

Send source assets, visual references, target duration and aspect ratio, audio needs, GPU model and VRAM, OS, ComfyUI version, deployment choice, deadline, and anything that must remain consistent.

A revision adjusts the agreed workflow or settings within written scope. A new model, new concept, extra stage, additional output, custom node, or changed deployment is a scope change and needs a custom quote.