# Stable Diffusion AI: images and data

## Files and results

Image-input tools accept JPG/JPEG, PNG, and WebP up to 10 MiB per file; a specific tool may impose additional dimension limits. Text-to-image does not need an upload. The Generator, Image to Image, and Photo Editor download ordinary results as JPG. Background Remover offers a transparent PNG cutout and PNG composites with chosen backgrounds. Output dimensions may differ from the source; keep your own copy of important files.

## Creations and sharing

Guest Recent is tied to the device used and is not a long-term backup. Signed-in Creations can be reopened across devices. Account Creations are not automatically put in a public gallery. A user may create an unlisted share link accessible to anyone who has the URL, then stop sharing. A saved Creation can be deleted from the library. See [FAQ](/faq) for workflow details.

Uploaded images are used to carry out the selected tool. For information handling, retention, and rights, consult the [Privacy Policy](/privacy) and [Terms](/terms-and-conditions). Do not infer an unstated training-use commitment or a universal retention period.

## Read next

[Site index](index.md) · [Features](features.md) · [Pricing](/pricing) · [Full overview](../llms-full.txt) · [llms.txt](../llms.txt)