- Meta AI’s image generator now runs on Muse Image, Meta’s own model launched in July 2026, replacing the previous third-party system.
- You can use it in the Meta AI app, on meta.ai, in Instagram Stories (U.S. only for now), and in WhatsApp (limited countries)—all for free.
- It’s agentic: Meta AI can search the web, write code and refine images to get details right, instead of producing one output and stopping.
- What you type and upload goes to Meta; don’t upload other people’s photos without permission.
Meta AI’s image generator lets you make pictures by typing a description. Since 7 July 2026, it runs on Muse Image, Meta’s first in-house image model, instead of relying on a third party. Muse Image is built into the Meta AI assistant—the same one that answers questions and helps with everyday tasks—so you can make pictures without leaving WhatsApp, Instagram or the Meta AI app.
This guide shows you exactly how to use the meta ai image generator, where it’s available right now, what makes Muse Image different, and what to know about privacy before you start.
Where you can use the Meta AI image generator
Meta AI with Muse Image is available in these places today:
- Meta AI app and meta.ai — Download the Meta AI app (iOS and Android) or visit meta.ai in your browser. Anyone can use it here.
- Instagram Stories — In the United States, you can type a prompt in Stories and Meta AI will generate an image. This feature is not yet available in other countries.
- WhatsApp — Meta AI is built into WhatsApp chats in a limited set of countries. Open a chat with Meta AI or type a prompt in any conversation to generate an image.
Meta has said that Facebook, Messenger, Instagram (beyond Stories) and WhatsApp will get broader rollout later. Right now, availability depends on where you live and which app you’re using.
How to make an image with Meta AI
The process is the same across all platforms. You type a description, and Meta AI produces an image. Here’s how.
In the Meta AI app or on meta.ai
- Open the Meta AI app or go to meta.ai in your browser.
- Type your prompt in the chat box. Be specific: instead of “a dog,” try “a golden retriever puppy sitting in a field of sunflowers at sunset.”
- Send the message. Meta AI will generate an image in a few seconds.
- If you want changes, describe them: “make the sky darker” or “add a red collar.” Meta AI will refine the image.
In Instagram Stories (U.S. only)
- Open Instagram and swipe to create a new Story.
- Tap the Meta AI icon or type a prompt in the text box.
- Meta AI will generate an image based on your description.
- You can edit, add stickers or text, then share it to your Story.
In WhatsApp
- Open WhatsApp and start a chat with Meta AI, or type in any existing chat.
- Type your prompt. You can also ask Meta AI to change the image after it’s generated.
- The image appears in the chat. You can download it or forward it.
Meta AI’s agentic approach means it doesn’t just generate one image and stop. It can search the web for reference information, write code to adjust details, and iterate on the output to get closer to what you asked for. This is different from older image generators that produce a single result per prompt.
What makes Muse Image different
Before July 2026, Meta AI used a third-party image model. Muse Image is Meta’s own model, built from the ground up. According to Meta’s announcement, Muse Image is designed to work as part of an agentic system: it can refine its own output, use web search to improve accuracy, and even write code to fix details like lighting or composition.
In plain terms, agentic means the AI doesn’t just guess once—it can take steps to improve the result. If your prompt asks for a “realistic photo of a cat wearing a red hat,” and the first attempt gets the hat color wrong, Meta AI can recognize the mistake and generate a corrected version without you having to ask again.
This iterative approach is unusual. Most image generators—like those covered in our guide to free AI image generators—produce one image per prompt and require you to manually rewrite the prompt or adjust settings if the output isn’t right. Muse Image’s ability to refine autonomously can save time, especially for users who aren’t practiced at writing detailed prompts.
Privacy and what Meta does with your prompts
When you use Meta AI’s image generator, your prompts and any images you upload are sent to Meta’s servers. Meta uses this data to generate your image, improve the model, and personalize your experience. According to Meta’s privacy policy, data from your interactions with Meta AI may be used for training and product development.
Here’s what you should know:
- Your prompts are not private — Assume that anything you type can be reviewed by Meta’s systems and, in some cases, by human reviewers for quality control.
- Don’t upload other people’s photos without permission — If you ask Meta AI to edit or generate variations of a photo, make sure you have the right to use that image. Uploading someone else’s photo without their consent may violate their privacy or Meta’s terms of service.
- Generated images are yours, but Meta’s terms apply — You own the images you create, but Meta’s terms allow the company to use, display and distribute them as part of its services. Read the full terms at meta.ai before sharing generated images publicly.
If privacy is a concern, consider using an image generator that runs locally on your device or one that doesn’t store your data. Our comparison of the best AI image generators includes privacy details for each tool.
Tips for better prompts
The more specific your prompt, the closer the result will be to what you want. Here are a few ways to improve your prompts:
- Describe the subject in detail — Instead of “a car,” try “a red 1960s convertible parked on a cobblestone street.”
- Specify the style — Add “watercolor painting,” “pencil sketch,” “photorealistic,” or “3D render” to guide the look.
- Include lighting and mood — “Golden hour light,” “dramatic shadows,” or “soft morning glow” can change the feel of the image.
- Mention the setting — “In a forest,” “on a beach at sunset,” or “inside a modern kitchen” gives context.
- Ask for changes one at a time — If the first image is close but not perfect, ask for one change: “make the sky darker” or “remove the background.” This works better than rewriting the entire prompt.
Because Meta AI is agentic, it can often fix small mistakes on its own. But clear prompts reduce the number of iterations needed.
How Meta AI compares to other image generators
Meta AI is not the only free image generator available. Tools like DALL·E, Midjourney, Stable Diffusion and others have been around longer and are widely used. Here’s how Meta AI fits in.
| Tool | Cost | Where you use it | Key feature |
|---|---|---|---|
| Meta AI (Muse Image) | Free | Meta AI app, meta.ai, WhatsApp, Instagram | Agentic refinement, built into apps you already use |
| ChatGPT (Images 2.5) | Free tier available; Plus, Pro for more | ChatGPT app, web | Integrated with ChatGPT’s text assistant; see how to use the ChatGPT image generator |
| Midjourney | Paid ($10/month and up) | Discord, web | High-quality, artistic results; see our Midjourney review |
| Google Gemini | Free tier; Advanced, Pro, Ultra for more | Gemini app, web | Integrated with Google’s AI assistant; see Google AI plans explained |
Meta AI’s main advantage is convenience: if you already use WhatsApp or Instagram, you don’t need to install another app or create a new account. The agentic refinement feature also sets it apart—most other tools require you to manually iterate on prompts.
However, leaderboards and independent comparisons change often, and no single tool is best for every use. If you need high-resolution outputs, commercial licensing or advanced control over style, you may prefer a dedicated tool like Midjourney or DALL·E. For quick, free images inside apps you already use, Meta AI is hard to beat.
Limits and what you can’t do
Meta AI’s image generator has some restrictions:
- No harmful, illegal or adult content — Meta’s content policy blocks prompts that request violence, explicit material, illegal activity or hate speech.
- No deepfakes or impersonation — You cannot generate images of real people (public figures or private individuals) in misleading contexts. Deepfakes are both a privacy risk and a trust issue; see deepfakes: the threat and how to detect them for more.
- No commercial use without checking terms — While you own the images you create, Meta’s terms place some restrictions on how you can use them. Read the terms at meta.ai before using generated images in ads, products or other commercial work.
- Output resolution may be lower than dedicated tools — Meta AI is optimized for social sharing, not print or large-scale commercial use. If you need high-resolution outputs, consider tools designed for that purpose.
When Meta AI image generator will expand
Right now, Muse Image is available in the Meta AI app, on meta.ai, in Instagram Stories (U.S. only) and in WhatsApp (limited countries). Meta has said that Facebook, Messenger and the full Instagram app will get the feature more broadly, but no specific dates have been announced. Availability depends on your location and the app you’re using.
If you don’t see Meta AI in your app yet, check back in a few weeks. Meta typically rolls out new features in stages, starting with a few countries and expanding over time.
Alternatives if Meta AI isn’t available yet
If Meta AI’s image generator isn’t available in your country or app, you have plenty of other options. Here are a few free and paid alternatives:
- ChatGPT — OpenAI’s assistant includes an image generator (Images 2.5) in the free tier. See our comparison of ChatGPT Free vs Go vs Plus vs Pro to understand what each plan includes.
- Google Gemini — Google’s AI assistant can generate images if you have Gemini Advanced, Pro or Ultra. See Google AI plans explained for details.
- Bing Image Creator — Microsoft’s free tool, powered by DALL·E, works in any browser and doesn’t require a Microsoft account.
- Craiyon — A free, no-signup image generator with lower resolution but unlimited use.
For a full list, see our guide to free AI image generators.
Frequently asked questions
Is Meta AI’s image generator free?
Yes. You can use the meta ai image generator at no cost in the Meta AI app, on meta.ai, and in WhatsApp and Instagram where it’s available. There are no daily limits mentioned in Meta’s public documentation, though Meta reserves the right to introduce limits in the future. No subscription or payment is required to generate images.
What is Muse Image and why did Meta switch to it?
Muse Image is Meta’s own image generation model, announced on 7 July 2026. Before that, Meta AI used a third-party model. Muse Image is designed to be agentic—it can search the web, write code and refine its own output to fix mistakes or improve details. This gives Meta more control over quality, speed and privacy, and allows the image generator to improve iteratively instead of producing one static result per prompt.
Can I use Meta AI’s generated images for commercial purposes?
You own the images you create, but Meta’s terms of service apply. Those terms grant Meta a license to use, display and distribute your generated images as part of its services. Before using generated images in ads, products or other commercial work, read the full terms at meta.ai. If you need images with clear commercial licensing, consider dedicated tools like Midjourney or Adobe Firefly, which offer explicit commercial licenses.
Why can’t I see Meta AI’s image generator in my Instagram or WhatsApp?
Meta is rolling out Muse Image in stages. Instagram Stories only has it in the United States right now, and WhatsApp has it in a limited set of countries. Facebook, Messenger and the full Instagram app will get broader access later. If you don’t see the feature yet, it hasn’t reached your location or app version. Check back in a few weeks, or use the Meta AI app or meta.ai website, which are available more widely.
Does Meta AI keep my prompts and generated images?
Yes. When you use Meta AI, your prompts and any images you upload are sent to Meta’s servers. Meta uses this data to generate your image, improve the model and personalize your experience. Data from your interactions may be used for training and product development. If privacy is a concern, consider using a tool that runs locally or one that doesn’t store your data. See our guide to the best AI image generators for privacy-focused options.
How does Meta AI’s image generator compare to ChatGPT’s?
Both are free and built into AI assistants. Meta AI’s advantage is that it’s inside WhatsApp and Instagram—apps you may already use daily—so you don’t need to switch apps or sign up for a new account. It’s also agentic, meaning it can refine images autonomously. ChatGPT’s image generator (Images 2.5) is more widely available globally and integrates with ChatGPT’s powerful text assistant, which may be better if you need both conversation and image generation in one place. See how to use the ChatGPT image generator for step-by-step instructions, and compare plans at ChatGPT Free vs Go vs Plus vs Pro.