Save 4 hours! We auto-setup your workflow! Free!

Drop your workflow.json — we handle every dependency, custom node, and model. Just open the link and run.

Auto-Setup Workflow Json (Free) Now!
ComfyUI > Nodes > Seedream4.5-comfyui-doubao4.5-comfyui > 🚀 Seedream 4.5 Image2Image

ComfyUI Node: 🚀 Seedream 4.5 Image2Image

Class Name

SeedreamImageToImage

Category
Seedream
Author
ninepoin4 (Account age: 1356days)
Extension
Seedream4.5-comfyui-doubao4.5-comfyui
Latest Updated
2026-07-04
Github Stars
0.0K

How to Install Seedream4.5-comfyui-doubao4.5-comfyui

Install this extension via the ComfyUI Manager by searching for Seedream4.5-comfyui-doubao4.5-comfyui
  • 1. Click the Manager button in the main menu
  • 2. Select Custom Nodes Manager button
  • 3. Enter Seedream4.5-comfyui-doubao4.5-comfyui in the search bar
After installation, click the Restart button to restart ComfyUI. Then, manually refresh your browser to clear the cache and access the updated list of nodes.

Visit ComfyUI Online for ready-to-use ComfyUI environment

  • Free trial available
  • 16GB VRAM to 80GB VRAM GPU machines
  • 400+ preloaded models/nodes
  • Freedom to upload custom models/nodes
  • 200+ ready-to-run workflows
  • 100% private workspace with up to 200GB storage
  • Dedicated Support

Run ComfyUI Online

🚀 Seedream 4.5 Image2Image Description

Generates images from up to four references and a prompt via Seedream, with batch output.

🚀 Seedream 4.5 Image2Image:

SeedreamImageToImage creates new images from one or more reference images and a written editing direction. It sends the references and instructions to a Seedream image-generation service, then downloads the results for use in your ComfyUI workflow. You can use it to explore visual changes while giving the service existing imagery to work from. The node can also request sequential generation when you need several results from one run. The changes you see depend on the service’s response; the node does not edit the reference images locally.

🚀 Seedream 4.5 Image2Image Input Parameters:

image

The required first reference image. The node converts it to a PNG and sends it to the service, so choose an image that clearly shows what you want to retain or change. It uses only the first image in an incoming batch. RGB images are accepted; for RGBA images, the transparency channel is discarded. There is no default value or stated size range.

prompt

The written instruction sent to the image-generation service. Describe the desired change and, where useful, what should remain recognizable. For example, you could ask to change a garment’s material while preserving the pose. The default is 保持模特姿势不变,改变材质... (“Keep the model’s pose unchanged, change the material...”); the field supports multiple lines. The source defines no character limit.

api_key

The credential used to authorize the request to the image-generation service. Without a valid key, the service may reject the request. The field is displayed as a password field and defaults to an empty string. The source defines no minimum or maximum length.

model

The model identifier sent with the request. Its default is doubao-seedream-4-5-251128. This is a free-text field, with no options or length limits defined in the node; the code does not check whether an entered identifier is available to your account.

size

The requested output-size setting sent to the service. The available options are 1K, 2K, 720x1280, and 1280x720; the default is 2K. Choose a portrait or landscape dimension when that shape suits your composition, or a 1K or 2K setting when you want to request one of those size presets. The actual result depends on what the service accepts and returns.

watermark

Controls the watermark setting sent to the service. The default is False. Set it to True when you want to request a watermark; the node itself does not add or remove one after downloading the image. This is a Boolean setting with True and False values.

timeout

The maximum time, in seconds, allowed for each individual network request: both the generation request and each result-image download use this setting. The default is 30, the minimum is 1, and the maximum is 300, in steps of 1. A longer timeout can help when a request or download takes more time, but it does not guarantee that the service will complete successfully.

sequential_image_generation

Controls the sequential-generation mode sent to the service. Options are disabled, auto, and enabled; the default is disabled. When it is disabled, the node does not send a sequential-generation image-count option. With auto or enabled, it sends that option along with the selected max_images value. The number of images ultimately returned is determined by the service response.

max_images

The requested maximum image count for sequential generation. The default is 1, the minimum is 1, and the maximum is 8, in steps of 1. This value is sent only when sequential_image_generation is auto or enabled; changing it while the mode is disabled has no effect on the request.

image2

An optional second reference image. Connect it when another view or visual reference would help explain your intended result. Like the required image, it is sent as a PNG, only the first image of an incoming batch is used, and any transparency is discarded. There is no default image or stated size range.

image3

An optional third reference image, sent alongside the references already connected. Use it when you have another relevant visual example; avoid adding references that conflict with your prompt. Only its first batch image is used, and any transparency is discarded. There is no default image or stated size range.

image4

An optional fourth reference image, bringing the maximum number of reference-image sockets to four. It is converted and sent in the same way as the other images: only its first batch image is used, and any transparency is discarded. There is no default image or stated size range.

🚀 Seedream 4.5 Image2Image Output Parameters:

image

The sole output is a ComfyUI IMAGE value containing the result downloaded from the service. If the service returns multiple images, the node combines them into one image batch on this same socket; it does not create extra output sockets. Downloaded images are converted to RGB, so their output does not retain transparency. If an exception occurs during preparation, the request, or result processing, the node prints an error and returns the original required image instead. An output image therefore does not, by itself, confirm that generation succeeded.

🚀 Seedream 4.5 Image2Image Usage Tips:

  • Start with one clear reference image and a prompt that names both the intended change and the elements you want preserved. Add image2 through image4 only when they contribute useful visual context.
  • Use 720x1280 for a portrait-shaped request or 1280x720 for a landscape-shaped request. If a size is rejected, try another listed option supported by your service configuration.
  • Leave sequential_image_generation at disabled and max_images at 1 when you need a straightforward single-result workflow. To request several results, choose auto or enabled and set max_images between 1 and 8; check the output batch for the number actually returned.
  • Increase timeout from its 30-second default if generation or image downloads repeatedly time out. The setting applies separately to the request and to each download.
  • If you connect a batch of references, remember that the node reads only the first image from each connected socket. Connect distinct images to the optional sockets when you need multiple references.
  • Check the node’s error output when the result looks identical to your first reference: on failure, the node returns that original image.

🚀 Seedream 4.5 Image2Image Common Errors and Solutions:

Unsupported number of channels in image 1: …, expected 3 (RGB) or 4 (RGBA)

  • Explanation: A connected reference image does not have the three RGB channels or four RGBA channels this node accepts. The image number in the message identifies the affected reference socket.
  • Solution: Convert that reference to an RGB or RGBA image before connecting it. For RGBA images, be aware that the node discards transparency.

API response missing 'data' field or empty

  • Explanation: The request received a response, but it contained no usable image entries in the data field the node expects.
  • Solution: Check that your key, model, size, and other request settings are accepted by the service. Review any available service response or logs for more detail, then retry.

Image URL is empty

  • Explanation: A returned image entry has no URL for the node to download.
  • Solution: Retry the request. If it persists, check the service response or logs; the node requires a downloadable URL for each returned image.

'url'

  • Explanation: This may appear when a returned image entry lacks a url field altogether. The node expects that field because it requests results in URL format.
  • Solution: Inspect the service response or logs and confirm that your request produces URL-based image results. If the response format differs, this node cannot process it as written.

401 Client Error or 403 Client Error

  • Explanation: These are possible authorization-related HTTP errors from the service. An empty, incorrect, or unauthorized API key may prevent generation.
  • Solution: Check the api_key value and confirm that it is authorized for the selected model. Do not share the key when asking for troubleshooting help.

400 Client Error

  • Explanation: The service may reject a request setting or its combination with other settings; the precise reason depends on its response.
  • Solution: Recheck the prompt, model identifier, size, and sequential-generation settings. Try the node defaults, then change one setting at a time to identify what is rejected.

Read timed out or ConnectTimeout

  • Explanation: The generation request or an image download did not complete within the configured timeout. The error may come from either network step.
  • Solution: Check your connection and try increasing timeout toward its 300-second maximum. If timeouts continue, check service availability and try again later.

ConnectionError

  • Explanation: The node could not establish or maintain a connection for the request or a result-image download.
  • Solution: Check your internet connection, network restrictions, and access to the service. Retry once connectivity is restored.

Error in Seedream ImageToImage Node: …

  • Explanation: This is the prefix the node prints for any exception it catches. The text after the colon gives the specific problem. After printing it, the node returns the original required image.
  • Solution: Read the full message following the colon and address that cause first. If the output appears unchanged, check this error line before revising your creative prompt.

🚀 Seedream 4.5 Image2Image Related Nodes

Go back to the extension to check out more related nodes.
Seedream4.5-comfyui-doubao4.5-comfyui
RunComfy
Copyright 2025 RunComfy. All Rights Reserved.

RunComfy is the premier ComfyUI platform, offering ComfyUI online environment and services, along with ComfyUI workflows featuring stunning visuals. RunComfy also provides AI Models, enabling artists to harness the latest AI tools to create incredible art.

🚀 Seedream 4.5 Image2Image