
Social creators and storytellers
Restore old photos, create sticker sheets, reimagine portraits, and generate shareable visuals for Instagram, Stories, and WhatsApp.
Muse Image is Meta's most advanced image generation model — an agent that searches, codes, self-refines, and edits with precision. Built for your world across Meta AI, Instagram, and WhatsApp.
Muse Image
Generated image preview
Your GPT Image 2 result will appear here
Muse Image moves beyond text-to-pixels — it works like an AI agent with tools, reasoning, and emergent self-refinement for production-quality results.
Instead of mapping prompts directly to pixels, Muse Image operates as an agent — invoking search and coding tools, reasoning at inference time, and improving its own outputs.

Edit images with surgical accuracy — change exactly what you ask for while maintaining coherence across iterative refinement and open-ended brainstorming.

Compose people, objects, clothing, styles, and environments from many reference images inline in your prompt for complex creative layouts.

Muse Image learns to search the web for factual and real-time visual references — improving accuracy on knowledge-intensive and current-events prompts.

Create with friends, reimagine Instagram photos, and generate dynamic content across Meta AI app, meta.ai, Instagram Stories, WhatsApp, and Facebook.
Use the image generator below to experiment with AI image creation inspired by Muse Image's agentic workflow.
Scroll to the generator and choose Text to Image or Image to Image — describe your subject, style, lighting, and composition.

Muse Image is designed for creators, marketers, and teams who need agentic control, social context, and production-ready image quality.

Restore old photos, create sticker sheets, reimagine portraits, and generate shareable visuals for Instagram, Stories, and WhatsApp.

Produce product shots, campaign key visuals, and on-brand compositions with readable layouts and consistent creative direction.

Prototype styles, test multi-reference compositions, and iterate edits with precision before final production workflows.
Experiment with our image generator now — while Muse Image rolls out across Meta AI with agentic tools, search, and Content Seal provenance.
Key strengths from Meta Superintelligence Labs' Muse Image launch and research.
Muse Image writes and executes code for accurate plots, QR codes, and rendered figures — and searches the web to ground generations in real-world facts.
The model reflects on and improves its own work — local edits when details are off, or full regenerations when larger parts need fixing.
Like language models, Muse Image improves with more inference-time reasoning — deliberate reasoning scales better than naive best-of-N sampling.
Muse Image ranks No. 2 on Arena for text-to-image, single-image editing, and multi-image editing by human-preference Elo at launch.
Muse Image integrates with Muse Spark for joint planning, animated GIFs, embedded websites, and interactive visual games.
Images created in Meta AI carry Content Seal — an invisible provenance signal that survives crop, compression, resize, and screenshots.
Muse Image sits alongside Muse Video, Muse Spark, and Meta AI — one connected stack for agentic image and video generation.

Meta's agentic image model with tool use, search, self-refinement, precision editing, and multi-reference composition — available today.

Preview text-to-video with exceptional visual fidelity and native audio — coming soon to creators and Meta AI.

Create, edit, and reimagine media with friends across Meta AI app, meta.ai, Instagram, WhatsApp, and Facebook.

Integrated with Muse Image for joint planning, code execution, and animated media — powering agentic creative workflows.
Read Meta Superintelligence Labs' research announcement and Meta's product launch for Muse Image.
Still have questions? We're here to help.