- How is this different from Canva's Magic Edit?
- Same idea — brush a region, type what should be there, AI generates it. Canva's Magic Edit is paid (part of Canva Pro) and limited per month. This tool is free with a small daily rate limit, and outputs at the original resolution of your photo. The underlying model (our AI) is more recent than older models.
- What can I magic-edit in a photo?
- Almost anything: add a hat to someone, change a shirt's colour, swap food on a plate, add a flower bouquet, replace a sky, remove a passing person, redecorate a room. The brush defines where; the prompt defines what. If you can describe it, you can usually generate it.
- How specific should my prompt be?
- Specific. "Roses" produces generic roses; "a small bouquet of pale pink roses in a clear glass vase, soft window light from the left" produces what you actually pictured. Include colour, lighting, and any key details. The AI interprets the prompt as a description of the brushed region only.
- Why is my edit subtly off?
- Magic edit works best when the brush fully covers the area to change (including any shadow or halo around it) and the prompt describes the new content in full sentences. If a result is off, refine the prompt or run again — every run is a different sample.
- Does the output have a watermark?
- No watermark, original resolution PNG. the AI processes internally at up to ~1024px on the long edge; we upscale the result back to your input size with Lanczos.
