ComfyUI Node

MiniMax Hailuo 02

The earlier MiniMax video tier that still earns its spot

By Runware·Created 2 years ago·Updated about a month ago· 140
MiniMax Hailuo 02
  • frameImages
  • video
positivePrompt
size720p (16:9)
durationfalse
duration_value0.00
numberResults1
providerSettings.minimax.promptOptimizerfalse
safetyfalse
safety.checkContentfalse
safety.modefast
ttlfalse
ttl_value60
outputFormatMP4
outputQuality95

If you're browsing the Runware/Video/minimax folder and wondering why there are three Hailuo nodes, here's the family tree: Hailuo 02 is the earlier generation, Hailuo 2.3 is the current flagship, and 2.3 Fast is the quick image-to-video lane. This node is the "previous model" tier - not a relic, but the cheaper, lighter entry point into MiniMax's video generation, and the one you reach for when you don't need the newest generation's fidelity for every clip.

The pitch is identical to its siblings: MiniMax Hailuo is a closed video model famous for physically believable motion, and this node runs it in the cloud through the Runware pack - no local GPU, no VRAM math, no downloading a checkpoint that doesn't exist. Where it differs from 2.3 is mainly muscle. 02 is older and a notch below on motion quality and prompt fidelity, but for rough animatics, quick drafts, or cost-sensitive work, it does the job at a friendlier price.

What you set

  • positivePrompt (required) - the motion-and-scene description.
  • frameImages - image-to-video input; feed a still and it animates from there.
  • width / height - 1024 default.
  • duration toggle + duration_value - seconds at the model's fps; off uses the default.
  • providerSettings.minimax.promptOptimizer - automatic prompt rewriting, and again: affects reproducibility. Off for stable seeds.
  • safety.mode - none/fast/full content checking.
  • outputFormat MP4/WEBM/MOV, outputQuality 20-99, numberResults 1-4.

Output is video (VIDEO).

How it works

Same pack machinery - model AIR + taskType over REST, media converted back to a native VIDEO, cost printed on the title bar when the run lands. Video is metered and per-run, and Hailuo runs aren't instant; expect a real queue. If the newer tier is overkill for what you're doing - testing a workflow, iterating on framing before committing to the good model - running 02 first is a genuinely reasonable cost optimization.

Installing

ComfyUI Manager → search Runware → install → restart. Manual:

cd ComfyUI/custom_nodes
git clone https://github.com/Runware/ComfyUI-Runware
pip install -r ComfyUI-Runware/requirements.txt

No model downloads; dependencies are runware-sdk, pillow, soundfile. API key via Settings → Runware API key, RUNWARE_API_KEY, or runware auth login.

Troubleshooting

The pack's universal first failure is the missing API key - set it before troubleshooting anything else. If your clips come out with worse motion than you remember from Hailuo demos, check you're actually on this node and not expecting 2.3's quality; that's the tier difference, not a bug. And keep promptOptimizer off if you care about reproducibility - the tooltip is explicit about it.

CategoryRunware/Video/minimax

Inputs (14)

NameTypeDefaultDescription
positivePromptSTRINGText prompt describing elements to include in the generated output.
frameImagesoptIMAGE
sizeoptCOMBO720p (16:9)Output size. Only one of resolution or dimensions is sent.
durationoptBOOLEANfalseEnable to set duration. Off uses the model's default.
duration_valueoptFLOAT0.00Length of the generated video in seconds. The total number of frames produced is determined by duration multiplied by the model's frame rate (fps).
numberResultsoptINT11–4Number of results to generate. Each result uses a different seed, producing variations of the same parameters.
providerSettings.minimax.promptOptimizeroptBOOLEANfalseEnable automatic prompt optimization. Affects reproducibility.
safetyoptBOOLEANfalseEnable to set safety. Off uses the model's default.
safety.checkContentoptBOOLEANfalseEnable or disable content safety checking.
safety.modeoptCOMBOfastSafety checking mode for video generation.
ttloptBOOLEANfalseEnable to set ttl. Off uses the model's default.
ttl_valueoptINT60Time-to-live (TTL) in seconds for generated content. Only applies when `outputType` is `URL`.
outputFormatoptCOMBOMP4File format for the generated video.
outputQualityoptINT9520–99Compression quality of the output. Higher values preserve quality but increase file size.

Outputs (1)

NameTypeDescription
videoVIDEO