Muse Image (https://www.meta.ai) generates and modifies images from natural language instructions, merging multiple photos and rendering text readable in the image.

Developed by Meta, the model relies on Muse Spark to analyze the instruction, search for information on the web if necessary, then compose the layout before generating the image. Available on meta.ai, in the Meta AI app, on WhatsApp, and in more than 30 Instagram Stories effects.

Presets (photo restoration, transformation into a figurine, 16-bit style, isometric diorama…) serve as a starting point.
Text management and persistence of objects and characters.

Warning: mentioning a public Instagram account in the instruction allows the person’s face, if it exists on the account, to be integrated into the generated image. This feature is enabled by default on public Instagram accounts: it can be disabled in the privacy settings related to AI reuse of content (see screenshot below).
Edit of 13/07: following the opposition of users to this possibility of using user content, Meta has removed this function.

Free for everyday use, with a generation cap not mentioned by Meta.