Meta has unveiled Muse Image, a new artificial intelligence-powered image generation model designed to transform simple ideas and written prompts into high-quality visual creations.
The technology, developed by Meta Superintelligence Labs, is the company’s first image generation model and is now available through Meta AI.
It is designed to serve as a creative partner, helping users generate, edit and share images across Meta’s platforms.
According to Meta, Muse Image combines advanced reasoning capabilities with image creation, allowing it to understand complex instructions, use contextual information and blend multiple visual references to produce realistic and customised images.
The company said the new model builds on the capabilities of Muse Spark, its large language model, by extending AI reasoning from text-based assistance to visual creativity.
“Meta AI is now able to help users turn their ideas into high-quality visuals that can be downloaded and shared directly to feeds, stories or chats,” the company said in a press statement yesterday
Meta said users can also request specialised creations, such as placing themselves in different locations, redesigning images, creating infographics or generating custom QR codes.
The platform also introduced suggested presets to help users begin the creative process.
Through these presets, users can restore old family photographs, experiment with different creative styles or transform images into unique digital designs.
Beyond personal creativity, Muse Image is expected to support practical applications, including interior design. Users can upload photos of their rooms and ask Meta AI to redesign spaces using style suggestions and products available online or through Facebook Marketplace.
Meta said Muse Image goes beyond basic image generation by analysing prompts before producing results. Working alongside Muse Spark, the system can plan layouts, combine multiple references and use relevant information to create visuals that better match users’ intentions.
The company also introduced features that allow users to tag Instagram accounts in the Meta AI app to include public images from specific profiles in their creations. Users, however, retain control over whether their content can be used for AI-generated images through privacy settings.
Editing has also been simplified, allowing users to make changes directly on generated images by sketching, marking areas for adjustment, or requesting further refinements without restarting the process.
