Nodes/BizyAirPlus/X AI Grok Imagine Image Edit (Channel Edition)
ComfyUI Node

X AI Grok Imagine Image Edit (Channel Edition)

Edit an image with Grok Imagine, right inside ComfyUI

By siliconflow·Created 5 months ago·Updated about a month ago· 20
X AI Grok Imagine Image Edit (Channel Edition)
  • images
  • api_config
  • image
  • request_id
  • urls
prompt
skip_errorfalse

The image-to-image sibling of the Grok T2I node. Same idea, one extra wire: you feed it an image you already have, describe the change, and Grok's model rewrites the picture to match. For the "channel edition" that means the request goes through SiliconFlow's BizyAir channel rather than xAI's own API - one SiliconFlow key, no separate xAI account.

Reach for this when your local img2img chain keeps giving you that uncanny thing you can't quite name. Grok's edit quality is genuinely different from a local SDXL or Flux edit at the same denoise - it tends to understand the instruction ("make her look 20 years older," "turn this daytime shot into neon-lit night") instead of just smearing pixels toward a prompt. That's the pitch for paying per generation: the model is doing the semantic lifting, not your sampler settings.

How it works

You can probably guess by now: this is a cloud API node, not a local pipeline. The images input (required, IMAGE) gets uploaded alongside the prompt, the BizyAir channel runs the edit, and the result comes back as an IMAGE tensor. Your GPU doesn't lift a finger. One thing to know up front: unlike the T2I variant, there's no aspect_ratio input here - the output inherits from the source image, which is how image editing should behave anyway.

The usual BizyAirPlus controls apply. api_config overrides base_url and api_key per request when you need a different endpoint or key. skip_error swaps a failed request for a placeholder instead of aborting the workflow - keep it on in batch jobs where one rejected image shouldn't torch the whole queue.

Inputs and outputs

  • images (required, IMAGE) - the photo you're editing.
  • prompt (required, multiline) - what to change. The more concrete, the better; Grok is unusually good at following detailed edit instructions.
  • Outputs: image (the edited result), plus request_id and urls strings for tracking and direct download.

Installing BizyAirPlus

Install the pack once and it covers every node in this family:

cd /path/to/ComfyUI/custom_nodes
git clone https://github.com/siliconflow/BizyAirPlus.git
cd BizyAirPlus
python -m pip install -r requirements.txt

or search "BizyAirPlus" in ComfyUI Manager. Restart, create an account at bizyair.ai, get an API key, and turn the BizyAirPlus action-bar button ON (Settings > BizyAirPlus > API Key, or the BIZYAIR_API_KEY env var). No models to download - the only dependencies are bizyair-cloudberry and bizytrd.

Common issues

Same three failure modes as every BizyAirPlus node. If the button never appears, your dependencies went into the wrong Python environment (pip show bizyair-cloudberry tells you). If cloud execution fails, check the key and your network, then read the ComfyUI console for the real error. And remember the bill - image editing per generation adds up fast if you're iterating. Set skip_error on, and maybe glance at the urls output once; it's useful when you want to double-check what actually got generated against the request_id in a support ticket.

CategoryBizyAirModelServices/BizyTRD/Grok Image

Inputs (4)

NameTypeDefaultDescription
imagesIMAGEPhoto
promptSTRINGPrompts
api_configoptBIZYAIR_OPENAPI_CONFIG单次请求覆盖base_url和api_key
skip_erroroptBOOLEANfalse开启后遇到错误不中断工作流,输出对应类型的错误占位符

Outputs (3)

NameTypeDescription
imageIMAGE
request_idSTRING
urlsSTRING