What are the two images used for?

The first (base) image is the photo you want to edit. The second (guide) image provides spatial or stylistic guidance—such as a mask, a layout mockup, or a reference that indicates where and how changes should happen. The ByteDanceSeedreamNodeV2 node conditions on both inputs along with your prompt.

How can I control which areas change while keeping the rest untouched?

Use a deliberate guide image. For precise localization, provide a clean mask‑style guide or a clear layout reference. Then use a focused prompt that describes only the intended change (e.g., “Change background to white; do not alter the subject”). Iterating with tighter guides generally increases control.

What image sizes work best?

Use reasonably high‑quality inputs and keep the base and guide images aligned in aspect ratio. If the node resizes automatically, feeding matched dimensions reduces warping and improves edge fidelity. For web use, a longest side around 1024–1536 px is a practical starting point.

Can I batch process multiple images with this setup?

This graph is configured for one base/guide pair at a time. ComfyUI can batch with additional nodes (e.g., directory loaders and iterators), but those are not included here. To batch, extend the workflow with batch‑friendly loaders and route them through the same ByteDanceSeedreamNodeV2 and SaveImageAdvanced pattern.

Seedream 5.0 Pro - ComfyUI Workflow

Seedream 5.0 Pro is an instruction‑guided image editing workflow built around the ByteDanceSeedreamNodeV2 model. It takes two images—a base image to be edited and a guide/reference image—plus a clear text prompt, then applies targeted changes while preserving the original lighting, depth, and surface texture. In ComfyUI, the images are brought in with LoadImage nodes, passed to ByteDanceSeedreamNodeV2 for editing, and written out using SaveImageAdvanced for organized, reproducible results.

Technically, ByteDanceSeedreamNodeV2 performs diffusion‑based edits conditioned on both the base image and your instruction prompt, with the second image acting as a guide for regions, layout, or stylistic cues. This dual‑conditioning lets you localize changes (e.g., background swaps, product color updates, layouted text) without destroying the scene’s geometry or material qualities. SaveImageAdvanced helps you version outputs with consistent filenames and metadata, making iteration straightforward.

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.

e2716c39cb46.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