img2img and Inpainting — The Underused Tools

Text-to-image is the tip of the iceberg

Text-to-image makes new pictures from nothing. img2img starts from a reference picture and transforms it — same composition, new style or subject. Inpainting lets you mask a specific area and regenerate only that area. Outpainting extends the image beyond its edges.

For editing existing images, these three are worth 10× more than raw text-to-image.

img2img

Denoising strength is the one dial that matters:

• 0.2–0.4 — subtle stylization, keeps the composition intact
• 0.5–0.7 — noticeable transformation, still recognizable
• 0.75+ — barely recognizable, effectively a new image

Use case: turn a rough sketch into a polished illustration. Turn a real photo into a painting. Restyle a product photo into a different mood.

Inpainting

Draw a mask over the part you want to change. The model regenerates only that region while keeping the rest pixel-perfect. Great for:

• Removing objects (mask them out, prompt "empty background")
• Swapping outfits on a character
• Fixing bad hands
• Changing an expression

Trick: for tiny fixes, mask slightly larger than the actual problem area — the model needs a bit of context to blend.

Outpainting

Extends the canvas beyond the original. The model has to invent what's outside the frame while matching the style of what's inside. Great for:

• Turning a portrait into a wider hero banner
• Adding vertical space above a product for text overlay
• Making a square image into 16:9

What Imagoat plans to ship

The AI Creative Tools section (in progress) will include one-click Background Removal, Background Replacement, Upscaling, Inpainting, Outpainting, and Image Variations — all wired to the same SDXL pipeline. Each tool costs 1–2 credits per image.