Nodes/Seedream4 Replicate/Seedream4 Replicate
ComfyUI Node

Seedream4 Replicate

A ComfyUI node in image/generation with 9 inputs and 1 output.

By Saganaki22·Created 11 months ago·Updated 11 months ago· 7
Seedream4 Replicate
  • image_input
  • images
◄api_key►
◄prompt►
◄size_preset2048x2048 (1:1)►
◄width2048►
◄height2048►
◄max_images1►
◄sequential_image_generationdisabled►
◄seed-1►
Categoryimage/generation

Inputs (9)

NameTypeDefaultDescription
api_keySTRINGYour Replicate API token (starts with r8_). Get this from replicate.com/account/api-tokens. Keep this secure and don't share it.
promptSTRINGDescribe what you want to generate. Be specific and detailed. Example: 'a photorealistic portrait of a woman with blue eyes, studio lighting, high detail'
size_presetCOMBO2048x2048 (1:1)Choose from common aspect ratios and sizes. Square formats work well for portraits and social media. Wide formats (16:9, 21:9) are good for landscapes. Use 'Custom' to set specific dimensions.
widthINT20481024–4096Image width in pixels (only used when size_preset is 'Custom'). Higher values = more detail but longer generation time. Must be between 1024-4096 pixels.
heightINT20481024–4096Image height in pixels (only used when size_preset is 'Custom'). Higher values = more detail but longer generation time. Must be between 1024-4096 pixels.
max_imagesINT11–15Maximum number of images to generate when sequential_image_generation is 'auto'. The AI decides how many to actually create (1 to this number). More images = higher cost.
sequential_image_generationCOMBOdisabledDisabled: Generate only 1 image. Auto: Let the AI decide if your prompt would benefit from multiple related images (like story sequences, character variations, etc.). Auto mode uses max_images as the limit.
seedINT-1-1–2147483647Random seed for reproducible results. Use -1 for random seed, or set a specific number to get the same image again. Same seed + same prompt = same result.
image_inputoptIMAGEOptional input image(s) for image-to-image generation. Connect an image here to use it as reference or starting point. Supports 1-10 images for multi-reference generation.

Outputs (1)

NameTypeDescription
imagesIMAGE—