Meta's first in-house AI image model — you can @-mention a public Instagram account and it pulls that person's public photos into the generated image.
Last verified · 2026-07-07 · by Moe Ameen
Muse Image is Meta's first in-house AI image generation model, launched on July 7, 2026. It is the second major release from Meta Superintelligence Labs, the group led by Alexandr Wang, after the Muse Spark language model earlier in the year, and was reportedly developed under the codename "Mango." It replaces the third-party and older systems Meta AI previously leaned on for image generation and is designed to power creative features across Meta's apps.
The standout feature is personal insertion. Inside Meta AI you can @-mention a public Instagram account, and Muse Image pulls that profile's public photos into the image it generates — so you can create a picture featuring a friend, a creator, or yourself without uploading anything. Meta pairs this with an opt-out control: being tagged and having your public images used is on by default, and Instagram users who don't want their content reused can turn it off in settings. The model leans on reasoning borrowed from Muse Spark to interpret complex prompts, blend multiple photos, plan layout, and pull in web context, and Meta highlights clean, legible text rendering inside images — historically a weak spot for image models.
At launch Muse Image is live in the Meta AI app and website, in WhatsApp (a limited rollout), and in Instagram Stories, where it powers more than 30 new AI effects and a set of preset prompts. You can also mark up a generated image directly to edit it. Meta says access is coming to Facebook, Messenger, and advertiser tools through Advantage+ creative. Every image carries an invisible watermark.
The honest framing: Muse Image is a consumer image generator that lives inside Meta's apps, not a content-production tool. It makes and edits stills, and its distribution is a single Instagram Story or a chat — it does not caption for other platforms, keep a brand voice across a week of posts, build carousels or blogs, or schedule and publish anywhere outside Meta's own surfaces. Meta also says a companion video model, Muse Video, is in development. Treat the specific effect counts, rollout, and availability as a launch-window snapshot.
Muse Image's trick — @-mention a public Instagram account and it drops that person into the scene — is genuinely new, but it produces a one-off still that lives inside a single Instagram Story or a WhatsApp chat, built from whoever you tagged. Two problems for a brand: the output is stranded on Meta's surfaces, and the "identity" in the image is a public account you don't own and that can opt out from under you at any time. Kompozy solves both. Bring a Muse Image still into Kompozy and it becomes the visual for a Photo Post, a brand-exact Carousel via HyperFrames, a Quote Graphic, or an Infographic Photo — captioned per platform and scheduled out to all nine connected channels plus blog and email, so the picture reaches TikTok, YouTube, LinkedIn, X, Pinterest, and Threads, not just a Meta feed.
Where Muse borrows a public account's face for a single image, Kompozy gives you a face-locked recurring identity you actually control: Gemini-powered Persona Photos and Persona Tweets keep the same avatar consistent across every post, and HeyGen-driven Persona Shorts and Persona Frames turn that identity into branded talking-head video. So the workflow is: use Muse Image for a striking, people-in-the-scene still when it fits, then let Kompozy fan one idea into 25–35 outputs across five buckets — held to one voice by your Persona Brief and banned-word filters — and publish the whole week from a single queue with Autopilot. Muse makes a picture inside Meta; Kompozy makes the content operation around it.
Muse Image is Meta's first in-house AI image generation model, launched July 7, 2026, from Meta Superintelligence Labs. It generates and edits images from conversational prompts inside Meta AI, WhatsApp, and Instagram, and its standout feature lets you @-mention a public Instagram account to pull that person's public photos into the generated image.
Inside Meta AI you @-mention a public Instagram account and Muse Image incorporates that profile's public photos into the image, so you can feature a friend, a creator, or yourself without uploading anything. Being tagged is on by default; Instagram users can opt out in settings if they don't want their content used.
Meta says Muse Image is free for everyday creation in the Meta AI app and site, WhatsApp DMs, and Instagram Stories, with additional usage available through Meta's paid AI subscription. Access for Facebook, Messenger, and advertisers via Advantage+ creative is rolling out.
No. Muse Image generates and edits stills inside Meta's own apps — its distribution is a single Instagram Story or a chat. It does not caption for other platforms, keep a brand voice, build carousels or blogs, or schedule posts elsewhere. A content engine like Kompozy handles captioning, brand governance, format fan-out, and cross-platform publishing.
Generate the image in Meta AI, then bring it into Kompozy. Kompozy makes it a Photo Post, Carousel, Quote Graphic, or Infographic, adds per-platform captions, fans the idea into text posts, a blog, and a newsletter under your Persona Brief, and schedules and publishes the set across nine platforms plus blog and email.