Vision AI
Liquid Prompt & OCR Studio
Qwen2.5-VL-3B · 4-bit
VRAM -- / -- GB
Drag & drop your image here
or · or press Ctrl+V to paste
JPG · PNG · WEBP · BMP · TIFF · AVIF
Analysis Mode
Flux / MJ Prompt
Max Tokens 512
Temperature 0.20
Top-P Sampling 0.90
Generated Output
Ready for inference
Upload an image and choose a mode to reverse-engineer prompts or transcribe text.
Batch & Dataset Configuration
Drag & drop multiple images or a whole dataset
or
0 files selected
Batch Status & Progress
# Image Status Generated Output Time Action
No batch job running yet. Select images or folder to begin.

Vision Analysis Presets

Pre-engineered instructions optimized for text-to-image prompts, camera aesthetics, OCR, and rich image captions.

Recent Generations

Browse recent single-image inferences, inspect full results, and re-copy generated text.

No previous generations found.