What do the optional stages do, and how do I enable them?

Two red ComfySwitchNode elements control optional branches: Prompt Enhancer (Optional) routes your text through MinimaxHailuo03ContextIRNode to add structured context/style before generation, and Upscale Video to 2K (Optional) uses a PrimitiveBoolean ("Upscale to 2K") to send the 768p result through an upscaling branch. Turn the relevant switch/boolean ON to include that stage.

Why must I generate at 768p to use the 2K upscale?

Per the author notes, the upscaler in this template only accepts 768p input. Keep the base model output at 768p when the “Upscale to 2K” flag is ON. Other base resolutions are not supported by the upscale branch.

How fast is the workflow and how long are the clips?

MiniMax H3 Max Turbo is optimized for rapid inference; the template typically returns a 768p video in about 1–3 seconds of generation time. The clips are short-form (about 5–15 seconds). Actual turnaround depends on network/API latency and your system.

What’s the difference between re-running and using the Regenerate node?

Re-running queues the full pipeline again from your current settings. MinimaxHailuo03RegenerateNode requests a fresh variant from the same prompt/context for quick iteration without editing inputs. Use it to explore more takes rapidly.

MiniMax H3 Turbo: Text to Video

MiniMax H3 Turbo: Text to Video is a lightweight ComfyUI template that turns a text prompt into a short video using the MinimaxHailuo03TextToVideoNode (MiniMax H3 Max Turbo). It requires no file inputs and produces a single 768p clip in roughly 1–3 seconds of model time, ideal for fast ideation, social-ready snippets, and quick A/B prompt testing. The core Create Video (Core) group feeds your prompt directly into the model, then SaveVideo writes the result to disk while SaveText logs the prompt and basic run info for reproducibility.

Two optional stages are controlled by red ComfySwitchNode elements. In Prompt Enhancer (Optional), the MinimaxHailuo03ContextIRNode can add structured context or style instructions to your prompt before generation. In Upscale Video to 2K (Optional), a PrimitiveBoolean flag ("Upscale to 2K") routes the 768p output through an upscaling branch. Per the author notes, the upscaler only accepts 768p input—so keep the base generation at 768p if you enable it. A MinimaxHailuo03RegenerateNode is also included to quickly request fresh variants without rewriting your prompt, making rapid iteration simple.

API

Use this workflow from code

Every Comfy workflow is a JSON graph. The payload below is this workflow, exactly as ComfyUI runs it — fetch it from the URL, keep it in version control, or load it in ComfyUI and run it node by node.

e001dfcb5ed0.json

Fetching workflow JSON…

Run it from TypeScript or Python with the Comfy SDK. The same code targets Comfy Cloud or a ComfyUI you host yourself — only the base URL changes.

// Install (beta)
npm i @comfyorg/sdk

// Run this workflow (TypeScript)
import { Comfy } from "@comfyorg/sdk";

const client = new Comfy({ apiKey: "comfyui-..." });
const wf = await client.workflows.fromFile("workflow_api.json");
const job = await client.run(wf);
await job.getOutputs("<output-node-id>")[0].toFile("output.png");

The SDK takes a workflow in API format: open this workflow in ComfyUI and use File → Export Workflow (API).

Comfy Cloud API access requires a plan with an API key.

SDK docs Get an API key

FAQ

Frequently Asked Questions