Reimagine an image with Grok and refinement
Generate an upscaled, refined reinterpretation of a reference image with Grok and Z-Image Turbo. Upload a reference image and set the style direction and detail instructions. QwenVL describes the reference; your style direction, that caption and your detail instructions form one positive prompt used by both generation and refinement. The uploaded image reaches Grok through its caption only, so identity, layout and details may change.
Workflow details
Workflow details
Runtime profile
- Publisher
- Sweet Tea Automation
- Status
- Listed
- Version
- Checking published version…
- Updated
- Oct 25, 2026
Hover nodes for Pipe controls
Core
Primary Pipe surface
Reference image
LoadImage
Style direction
StringConcatenate
Detail instructions
StringConcatenate
Image generation
GrokImageNode
Base image size
ImageScaleToTotalPixels
Final image scale
ImageScaleBy
Refinement
KSampler
Extended
Additional surfaced controls
Refinement exclusions
CLIPTextEncode
Sampling configuration
ModelSamplingAuraFlow
What you need
- Model files
- 0.3 GiB known subtotal
Remaining hardware or download-size estimates are not available yet.
Output
- Sweet Tea Preview Image · image
Requirements
Models
4- Found
VAE · 320 MB · SAFETENSORS · Hugging Face · Comfy-Org/z_image_turbo
Text encoder · 7.49 GB · SAFETENSORS · Hugging Face · Comfy-Org/z_image_turbo
- File unverified
Upscaler · Hugging Face · Comfy-Org/Real-ESRGAN_repackaged
Checkpoint · 11.5 GB · SAFETENSORS · bf16 · Hugging Face · Comfy-Org/z_image_turbo
Node packs
1Node pack · GitHub · tea-time-labs/sweet-tea-nodes
Sources
1 source
Sources
1 source