Uplaoded for my own sake.
Input an image and it will extend it further. Adjust the sliders for positioning.
Workflow based on AIFuzz's outpainting Youtube video.
Uses a custom vibe coded node to extract a prompt: https://github.com/Pochog/ComfyuiImageExtractPrompt
Uses Flux 1 Fill Dev. And Qwen VL LLM. Might need a bit of VRAM