ComfyUI Node

Ideogram 4.0 Remix

The hosted Ideogram 4 remix — no weights, no filter fights

By Runware·Created 2 years ago·Updated 17 days ago· 140
Ideogram 4.0 Remix
  • seedImage
  • advancedFeatures.watermark.image
  • image
positivePrompt
size2K (1:1)
strengthfalse
strength_value0.00
numberResults1
advancedFeaturesfalse
advancedFeatures.watermark.bgColor
advancedFeatures.watermark.displayPosition(default)
advancedFeatures.watermark.fontColor
advancedFeatures.watermark.opacityfalse
advancedFeatures.watermark.opacity_value0.10
advancedFeatures.watermark.text
settings.copyrightDetectionfalse
settings.renderingSpeedDEFAULT
safetyfalse
safety.checkContentfalse
ttlfalse
ttl_value60
outputFormatJPG
outputQuality95

Ideogram 4.0 is the release the community argued about for weeks - open weights in June 2026, but non-commercial, quantized-only, and with a safety filter trained into the model that made ComfyUI's lead developer publicly disclaim it wasn't a bug. This node is none of that. Runware_ideogram_4_0_remix is the hosted API version of the 4.0 remix, served through Runware. No weights, no license treadmill, no refusal-image-bytecode forensic debates. You get the 4.0 engine, you pay per call.

It's the remix variant of the current-generation model: give it a seed image and a prompt, and it transforms the original with 4.0's strength - which, notably for this family, includes its character and IP recall. People found that recall was the real standout of 4.0, and it applies here too.

How it works

Required: seedImage and positivePrompt. Width/height are optional this time (defaulting to 1024) rather than required - one less thing to trip over.

The inputs that actually matter:

  • strength (boolean) + strength_value (0 to 1, step 0.01) - this is the pair to understand. strength defaults to false, which means "let the model pick." Flip it on and strength_value overrides the automatic choice. Lower values keep more of the original; higher values drift further. That "off = model default" gating pattern is used all over this pack, so once you've seen it here you've seen it everywhere.
  • settings.renderingSpeed - TURBO / DEFAULT / QUALITY. The classic speed-quality tradeoff. DEFAULT is a fine starting point.
  • settings.copyrightDetection - a 4.0-specific opt-in: post-generation checks using Hive's likeness and logo detectors. Worth flipping on for commercial-ish work, knowing the rest of the pipeline may still trip the moderation anyway.

There's no negativePrompt on this node - 4.0 dropped it, so don't go hunting. numberResults gives you up to 20 variations, each on its own seed.

Output is a native IMAGE, straight into Preview Image or Save Image.

Installing it

Same pack, same drill as the rest of the Runware nodes:

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

Restart, set your API key in ComfyUI Settings → Runware API key (or the RUNWARE_API_KEY env var). Light dependencies - runware-sdk, pillow, soundfile - and zero model downloads, which is the whole pitch.

Where people get burned

Two things. First, people confuse this with the open-weights 4.0 and expect to fiddle with the JSON prompt builder or fight the filter - you don't, because you're renting the hosted engine and its moderation runs server-side (safety.checkContent will surface as NSFW: yes/no in the title bar). Second, the strength/strength_value split confuses everyone once: leaving strength off isn't a bug, it's "auto," and the moment you flip it on your slider actually does something. And as always with this pack: no key means instant errors, and it's a paid top-up API - cheap per image, but mind the minimum top-up before you batch 20 results.

CategoryRunware/Image/ideogram

Inputs (22)

NameTypeDefaultDescription
seedImageIMAGE
positivePromptSTRINGText prompt describing elements to include in the generated output.
advancedFeatures.watermark.imageoptIMAGE
sizeoptCOMBO2K (1:1)Output size. Only one of resolution or dimensions is sent.
strengthoptBOOLEANfalseEnable to set strength. Off uses the model's default.
strength_valueoptFLOAT0.000–1Strength of the transformation. Lower values result in more influence from the original input. When omitted, the strength is chosen automatically from the prompt. Supplying a value overrides that automatic choice.
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.
settings.copyrightDetectionoptBOOLEANfalseOpt into post-generation copyright detection using Hive likeness and logo checks.
settings.renderingSpeedoptCOMBODEFAULTGeneration speed/quality tradeoff.
safetyoptBOOLEANfalseEnable to set safety. Off uses the model's default.
safety.checkContentoptBOOLEANfalseEnable or disable content safety checking. Increases total generation time.
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