MuseCodes

Guides / Getting started

What Is Muse? A Plain-English Guide to Meta’s AI Agent

Muse is Meta’s personal AI agent. Unlike a chatbot that only answers questions, Muse can carry out tasks for you, keep working after you close the app, and create images and video. This guide explains what it is in plain terms, so you can decide whether it’s worth signing up.

The short version

Meta announced Muse on September 8, 2026. It runs on Meta’s Muse Spark model, and each user gets their own agent running on a cloud computer that Meta hosts. You talk to it in the Muse app, on the web at muse.ai, or inside WhatsApp, and it can do multi-step jobs: research, writing, planning, and making images and video clips.

The easiest way to think about it: a chatbot gives you an answer, while an agent gives you a finished result. Ask a chatbot for a mover and you get tips on choosing one. Ask Muse and it can go and collect quotes for you.

What Muse can do

Run tasks for you

Muse can work through jobs that take several steps, like comparing options, filling in details, and reporting back. It can also run scheduled jobs on a repeating basis, and it keeps an activity log so you can see what it did and what it plans to do next.

Make images

Image generation is powered by Muse Image, which Meta launched in July 2026. It can create images from a written description, edit a photo you upload, combine several reference images into one scene, and render readable text such as signs and menus. Our image generation tutorial covers this step by step.

Make video

Meta announced Muse Video alongside Muse Image. It generates clips of up to 10 seconds with matching sound, and it’s currently in a limited beta. Longer videos are made by joining clips together; read what Muse Video is for the current status. See how to make videos with Muse for the workflow we use for full YouTube videos.

Remember your preferences

Muse remembers things you tell it, such as your writing style or the format you like, and you can ask it to forget specific details at any time.

Read your documents

You can hand Muse a document and it will work from it. In our experience it picks up both the text and the images inside a document, which is useful when you have a list of prompts with reference pictures. More on this in our consistent characters guide.

Where Muse is available

At the time of writing, Muse is available in the United States (since September 8, 2026), Canada (since September 18) and, according to reports, Japan (since October 23). Other countries don’t have dates yet; see where Muse works. It comes as:

You need to be 18 or over to subscribe to a paid plan. Availability is changing quickly, so check muse.ai for the current list of countries.

Is Muse free?

Yes, you can start for free. Everyone gets a usage allowance measured in tokens that refreshes regularly, and there are paid plans for heavier use. New users can also get a large one-time bonus by redeeming an invite code within 48 hours of signing up. Our tokens guide explains what tokens are and how fast different tasks use them.

What Muse is not good at (yet)

Muse is new, and early reviews point out some rough edges worth knowing about:

Who Muse is for

Muse is most useful if you have repeatable jobs you’d like to hand off, or if you create content and want images and short clips without switching between several tools. If you only want quick answers to questions, a regular chatbot will feel faster.

Next steps

Wondering how it stacks up against other assistants? Read Muse vs ChatGPT. Ready to try it? Follow our sign-up and invite code guide so you don’t miss the 48-hour bonus window, then grab a code from our invite code list.