ComfyUI Node

Ideogram 3.0

Character consistency plus 63 style presets in one node

By Runware·Created 2 years ago·Updated about a month ago· 140
Ideogram 3.0
  • advancedFeatures.watermark.image
  • providerSettings.ideogram.characterReferenceImages
  • providerSettings.ideogram.characterReferenceImagesMask
  • providerSettings.ideogram.styleReferenceImages
  • image
positivePrompt
negativePrompt
size1K (1:1)
seed0
upscaleFactor(default)
numberResults1
advancedFeaturesfalse
advancedFeatures.watermark.bgColor
advancedFeatures.watermark.displayPosition(default)
advancedFeatures.watermark.fontColor
advancedFeatures.watermark.opacityfalse
advancedFeatures.watermark.opacity_value0.10
advancedFeatures.watermark.text
providerSettings.ideogram.magicPromptAUTO
providerSettings.ideogram.renderingSpeedDEFAULT
safetyfalse
safety.checkContentfalse
providerSettings.ideogram.stylePreset(default)
providerSettings.ideogram.styleTypeAUTO
ttlfalse
ttl_value60
outputFormatJPG
outputQuality95

Ideogram 3.0 is the serious one. The 2a and 2a Remix nodes are for quick text-and-image jobs; this is the tier where Ideogram added real photo realism, style references, and - the headline feature - character consistency. That's the thing local checkpoints still fight with daily (the KB's identity-preservation topic exists precisely because nobody does it effortlessly), and here it's a couple of image sockets away.

It's a cloud node through the Runware/ComfyUI-Runware pack, so again: no local model, no GPU, native IMAGE out. But it's the most parameter-heavy Ideogram node in the pack, so it deserves a walkthrough.

How it works

The node calls Ideogram 3.0 through Runware's SDK. What makes it different from the 2.x nodes:

  • Character reference. providerSettings.ideogram.characterReferenceImages - feed it 1-3 images of a person or character, and the output keeps that identity. The optional characterReferenceImagesMask lets you tell it which pixels are the character, which tightens consistency a lot.
  • Style reference. providerSettings.ideogram.styleReferenceImages - transfer visual language from reference images.
  • Style presets. providerSettings.ideogram.stylePreset has 63 curated presets - ART_DECO, BAUHAUS, BLUEPRINT, 80S_ILLUSTRATION, AURA, and friends. This is where you stop describing an aesthetic in 40 words and just pick the drop-down.
  • Upscale. upscaleFactor of 1, 2, or 4 doubles/quadruples dimensions in the same call instead of sending you off to a separate upscale pass.

The inputs that matter

  • positivePrompt - required. The scene.
  • width / height - required, 1024 default.
  • providerSettings.ideogram.characterReferenceImages (+ optional ...Mask) - the consistency feature. The reason to pick this node over every other Ideogram node in the pack.
  • providerSettings.ideogram.stylePreset - the 63-preset dropdown. (default) if you want to stay uninfluenced.
  • providerSettings.ideogram.styleType - AUTO / GENERAL / REALISTIC / DESIGN / RENDER_3D / ANIME / FICTION.
  • upscaleFactor - do it here, don't buy a second pass.
  • negativePrompt, seed, providerSettings.ideogram.magicPrompt, renderingSpeed - familiar by now.

Installing

cd ComfyUI/custom_nodes
git clone https://github.com/Runware/ComfyUI-Runware
pip install -r ComfyUI-Runware/requirements.txt

Restart, or ComfyUI Manager → search Runware. API key via Settings, RUNWARE_API_KEY, or runware auth login.

Where people get burned

Character consistency is real but not magic - it's strongest when the reference images are the same character in consistent lighting, and the mask helps when they aren't. Don't skip the mask and then complain the face drifts on a complex group shot. Also, one thing to keep in mind from the KB: Ideogram 4.0 shipped open weights in June 2026 and then promptly lost the ecosystem race to Krea 2 over licensing, which tells you Ideogram's open future is uncertain - but these API nodes don't depend on that, they keep working. And since 3.0 is the priciest Ideogram tier here, the title-bar cost readout is worth actually glancing at.

CategoryRunware/Image/ideogram

Inputs (27)

NameTypeDefaultDescription
positivePromptSTRINGText prompt describing elements to include in the generated output.
advancedFeatures.watermark.imageoptIMAGE
providerSettings.ideogram.characterReferenceImagesoptIMAGE
providerSettings.ideogram.characterReferenceImagesMaskoptIMAGE
providerSettings.ideogram.styleReferenceImagesoptIMAGE
negativePromptoptSTRINGPrompt to guide what to exclude from generation. Ignored when guidance is disabled (CFGScale ≤ 1).
sizeoptCOMBO1K (1:1)Output size. Only one of resolution or dimensions is sent.
seedoptINT00–2147483647Random seed for reproducible generation. When not provided, a random seed is generated in the unsigned 32-bit range.
upscaleFactoroptCOMBO(default)Factor by which to upscale the generated image. A value of 2 doubles width and height.
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.ideogram.magicPromptoptCOMBOAUTOAutomatic prompt enhancement.
providerSettings.ideogram.renderingSpeedoptCOMBODEFAULTGeneration speed/quality tradeoff.
safetyoptBOOLEANfalseEnable to set safety. Off uses the model's default.
safety.checkContentoptBOOLEANfalseEnable or disable content safety checking.
providerSettings.ideogram.stylePresetoptCOMBO(default)Artistic style preset.
providerSettings.ideogram.styleTypeoptCOMBOAUTOVisual style and rendering approach.
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