ComfyUI Node

Seedream 5.0 Lite

ByteDance's image model, fast tier, sequence mode

By Runware·Created 2 years ago·Updated about a month ago· 140
Seedream 5.0 Lite
  • referenceImages
  • advancedFeatures.watermark.image
  • image
positivePrompt
size2K (1:1)
numberResults1
advancedFeaturesfalse
advancedFeatures.watermark.bgColor
advancedFeatures.watermark.displayPosition(default)
advancedFeatures.watermark.fontColor
advancedFeatures.watermark.opacityfalse
advancedFeatures.watermark.opacity_value0.10
advancedFeatures.watermark.text
providerSettings.bytedance.maxSequentialImagesfalse
providerSettings.bytedance.maxSequentialImages_value1
safetyfalse
safety.checkContentfalse
ttlfalse
ttl_value60
outputFormatJPG
outputQuality95

Seedream is ByteDance's flagship image model, and like Seedance on the video side, it's an API product - no open weights, so a node like this (bytedance:[email protected]) is how it gets into your ComfyUI graph at all. The "Lite" suffix is the speed/cheapness tier, same family tradeoff as the Seedance Fast node: lower cost, lower latency, slightly less polish than the full model. For drafting and iteration that's the right call, and it's why Lite exists in the catalog next to the full-fat versions.

The pack setup is the same as every other Runware node: install via ComfyUI Manager (search Runware) or git clone https://github.com/Runware/ComfyUI-Runware into custom_nodes/, add a Runware API key in ComfyUI Settings, and queue. What's interesting about this particular node isn't the basics - it's the sequence feature buried in the provider settings.

The feature worth knowing: sequential images

providerSettings.bytedance.maxSequentialImages is the sleeper input. Enable it and set a count (1–15) and the model generates a sequence of images in one request - same subject across multiple frames. Combined with referenceImages, that's a storyboard or a pose-sequence tool: feed a character reference, get N consistent images of that character in different poses or angles. The constraint is right there in the tooltip: reference images plus sequential images can't exceed 15 total, and the model may generate fewer than you asked. It's the closest this pack comes to "consistent character sheet" in a single node, and it's easy to miss because it's gated behind a provider-setting toggle.

Inputs that matter

  • positivePrompt (required) - subject + style + variation instructions. For sequences, describe the arc you want across the frames.
  • width / height (1024 default) - the basics.
  • referenceImages - conditioning for style/subject consistency, including across the sequential batch.
  • providerSettings.bytedance.maxSequentialImages - enable + set the count for multi-frame generation. The distinctive feature.
  • numberResults (1–20) - regular variations on top of any sequencing.
  • watermark socket, safety gate, outputFormat/outputQuality.

The image output is a native IMAGE; a sequence comes back as a batch tensor, so it flows into anything that handles batches (or split it and preview each frame).

Where people get burned

  • Missing the sequence feature. It's behind a providerSettings.-prefixed gate, which reads like plumbing you should ignore. It's not - it's the node's reason to exist.
  • The 15-total cap. Reference images count against the same pool as sequential images. Load 10 references and you're capped at 5 generated frames.
  • The toggle habit - maxSequentialImages is an "enable to set X" boolean with a _value field. Flip the toggle, set the count.
  • Lite is Lite. Drafts and iteration are where it belongs; when a prompt earns a final render, the non-Lite Seedream tier is the upgrade path.

Seedream 5.0 Lite is the node you'd pick when you want ByteDance's image quality without the price of the flagship, and the sequence mode is a genuinely unique capability in this pack. Draft in Lite, sequence your consistent batches, promote the keepers.

CategoryRunware/Image/bytedance

Inputs (20)

NameTypeDefaultDescription
positivePromptSTRINGText prompt describing elements to include in the generated output.
referenceImagesoptIMAGE
advancedFeatures.watermark.imageoptIMAGE
sizeoptCOMBO2K (1:1)Output size. Only one of resolution or dimensions is sent.
numberResultsoptINT11–20Number of results to generate. Each result uses a different seed, producing variations of the same parameters.
advancedFeaturesoptBOOLEANfalseEnable to set advancedFeatures. Off uses the model's default.
advancedFeatures.watermark.bgColoroptSTRINGBackground color in hex format.
advancedFeatures.watermark.displayPositionoptCOMBO(default)Watermark position.
advancedFeatures.watermark.fontColoroptSTRINGText color in hex format.
advancedFeatures.watermark.opacityoptBOOLEANfalseEnable to set advancedFeatures.watermark.opacity. Off uses the model's default.
advancedFeatures.watermark.opacity_valueoptFLOAT0.100.1–1Watermark opacity.
advancedFeatures.watermark.textoptSTRINGWatermark text.
providerSettings.bytedance.maxSequentialImagesoptBOOLEANfalseEnable to set providerSettings.bytedance.maxSequentialImages. Off uses the model's default.
providerSettings.bytedance.maxSequentialImages_valueoptINT11–15Maximum number of sequential images to generate. Combined total of reference images plus sequential images cannot exceed 15. The model may generate fewer images than requested depending on prompt complexity.
safetyoptBOOLEANfalseEnable to set safety. Off uses the model's default.
safety.checkContentoptBOOLEANfalseEnable or disable content safety checking.
ttloptBOOLEANfalseEnable to set ttl. Off uses the model's default.
ttl_valueoptINT60Time-to-live (TTL) in seconds for generated content. Only applies when `outputType` is `URL`.
outputFormatoptCOMBOJPGFile format for the generated image.
outputQualityoptINT9520–99Compression quality of the output. Higher values preserve quality but increase file size.

Outputs (1)

NameTypeDescription
imageIMAGE