Do I need an API key or internet access for HunyuanImageEditApi?

Yes. HunyuanImageEditApi sends your image and prompt to the HY Image 3.5 Preview service, so an internet connection is required. If the node exposes API credentials or an endpoint field, enter them as instructed by the node’s tooltip or your provider’s docs.

How should I write prompts for consistent, controlled edits?

Be explicit about both changes and constraints. Include "keep" statements (e.g., keep pose, angle, composition), specify materials/colors/lighting, and avoid conflicting directives. Start simple, verify the result, then incrementally add detail.

What image sizes work best?

Use a reasonably sized input (e.g., 1–4 MP) to balance detail and speed. Extremely large images increase upload time and may be resized by the API. If you see cropping or scaling artifacts, try resizing to a standard aspect and re-run.

How do I preserve subject identity and composition?

Anchor the edit by referencing what must remain unchanged (e.g., "keep the same bike shape and camera angle"). Avoid style prompts that imply a full scene replacement when you only want local changes. If the node offers a strength or preserve-structure control, keep it conservative to maintain the original layout.

HY Image 3.5 Preview: Image Edit

This workflow demonstrates prompt-guided image editing with the HY Image 3.5 Preview model using ComfyUI. It takes one source image, interprets a natural‑language instruction, and returns a single edited image. The core of the graph is the HunyuanImageEditApi node, which sends your image and prompt to the HY Image 3.5 Preview service for multimodal understanding and edit synthesis. LoadImage brings your source (for example, the provided futuristic_brutalist_motorcycle.png), and SaveImageAdvanced writes the result to disk with your preferred filename and format. A MarkdownNote in the canvas provides quick inline tips and links.

Technically, the workflow preserves the input image as a visual anchor while the HunyuanImageEditApi aligns your text prompt to the existing scene. This lets you direct high‑level changes (materials, colorways, backgrounds, lighting, mood) without manually masking or compositing. Because the inference runs through an API, you get consistent edits without managing local model weights, and SaveImageAdvanced gives you control over output naming, format, and quality for production handoff. It’s a practical setup for portrait cleanup, product/hero image variations, and fast turnarounds on marketing visuals.

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.

8911701d2fed.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