How to Use Muse Spaces: Build a Tracker by Describing It

Short answer: a space is a small custom tool inside the Muse app, such as a tracker, a planner or a project board, that Muse builds after you describe it in chat. Meta announced Spaces on 6 November 2026 and says they are available everywhere Muse is available. You make one by describing what you want, then keep adjusting it in conversation. Sharing a space with other people is not available yet.

This guide explains what a space is, gives you a first one to try, and covers the part that deserves the most care: which apps you let it read from. It sticks to what Meta and the press have reported as of 9 November 2026. Meta’s announcement does not name buttons or menus, so the steps below describe what to do rather than inventing screens.

What a space is, and which Muse this is

Meta now uses the Muse name for several things. Muse is the personal agent app. Muse Spark and Muse Wonder are models. Muse Code is a coding agent. Spaces belong to the first one, the app. If the model names are what confused you, what Muse Spark is sorts them out, and the Meta Muse app guide covers getting the app itself.

Meta describes a space as a custom experience you create by describing your idea, shaped around your routines, goals and interests. In plainer terms, it is a page in the app that stays put. A normal chat scrolls away. A space is something you come back to: a list of gift ideas with prices, a planner for the week, a board for a home project.

TechCrunch’s report says spaces live in the sidebar of the app, next to your conversations with the agent. Its reporter also admits the feature is a little hard to pin down from the outside, which is fair. The easiest way to understand one is to build one.

Who can use Spaces

Meta’s wording is that spaces are available everywhere Muse is available. Muse launched in the US on 8 September 2026, and Meta announced its launch in Japan on 23 October. On 9 November 2026 Apple’s lookup service returned no Muse from Meta listing for the German, French or Korean App Store, so if you are in one of those countries, Spaces is not something you can try yet.

Meta’s post does not say whether Spaces needs a paid plan, and it gives no separate price. Muse itself has a free plan with a usage meter and paid tiers, covered in Meta Muse pricing. Building and updating a space is still the agent doing work, so expect it to count against the same meter. That is an inference, not something Meta has stated.

How to make your first space

Meta’s own instructions are one line: tell Muse about it, and it builds it for you. In practice that means the quality of the space depends almost entirely on the quality of the description. Here is the order that works for most first attempts.

1. 01Update the Muse app first. A newly announced feature does not always reach every account on the same day.
2. 02Pick something you already track by hand, in a note, a spreadsheet or your head. You will know straight away whether the space is right.
3. 03Describe the purpose, the fields you want, and how it should be laid out. Say what it should not do.
4. 04Look at what Muse builds, then correct it in the same conversation. Meta says you can keep adjusting a space with Muse as you use it.
5. 05Decide where it lives. Meta says you can add a space to the sidebar, move it into the main tabs, hide it, or schedule when it appears.

Scheduling is the detail worth noticing. Meta’s example is a weekly planner that sits at the front during the working week and gives way to weekend plans on Saturday. If you build two or three spaces, that is what stops the app filling up with tools you only need on certain days.

A worked example: a weekly work planner

The request below was written by Coursium for this guide. It has not been run in Muse, so treat it as a template, not a tested result.

Build me a space for planning my working week. Columns: task, project, due day, status (not started, in progress, done) and a notes field. Group tasks by day, Monday to Friday. Show anything overdue at the top in its own section. Do not add tasks yourself, do not send messages or book anything, and do not connect any app until I ask.

Why it is written that way. It names the exact fields, so Muse does not guess. It says how to group and sort, which is where a vague request usually goes wrong. And it closes with what the agent must not do. Muse is an agent that can act on your behalf, which is the difference between it and a chatbot, explained in AI agents vs LLMs. A planner should not start doing things on its first day.

Then test it. Add three real tasks, mark one overdue, and see whether it moves to the top. If it does not, say exactly what you expected. One specific correction is worth more than rewriting the whole request.

Connecting apps: start with one, and only what it needs

The most useful part of Spaces is also the one to slow down for. Meta says that if you connect the apps you use, your spaces keep themselves up to date in the background. Its examples include school emails, a calendar, Instagram saves and Marketplace messages. Neither Meta nor TechCrunch lists which apps can be connected, so check that inside the app rather than trusting a third-party list.

Background updates mean the space reads your data when you are not watching. That is the point, and it is also the risk. A few habits keep it manageable:

• —Connect one source per space, and only the one it needs. A gift tracker does not need your work email.
• —Leave out anything you would not paste into a chat yourself: medical letters, payslips, other people’s private messages.
• —Read Meta’s Muse security page before connecting a work account, and check whether your employer allows it at all.
• —Disconnect a source when the space has done its job, such as a trip planner after the trip.

Meta’s announcement says spaces are personal from the start, but it gives no detail on retention or permissions specific to Spaces, and links only to its general privacy policy. The wider question of what Muse can see and do once connected is covered in is Meta Muse safe.

Check what the space shows you

A space that updates itself looks authoritative. A neat table with a status column feels like a fact. It is still AI output: an email can be misread, a date can land on the wrong day, a price can be stale. For anything that matters, open the source once and compare. The method for doing that when you are not an expert in the subject is in how to check an AI answer.

The first week is when this matters most. Once you have seen a space get the same kind of item right ten times in a row, you can check it less often. Until then, check it.

What is not there yet

• —Sharing. Meta says it is exploring ways to share spaces with others. It has given no date.
• —A list of supported apps. Meta describes connecting apps but names none.
• —Detailed instructions. Meta’s post has no step-by-step guide, so screens may differ from any description you read, including this one.

If you want to compare Spaces with what other agents do, AI agent examples shows the range, from booking tools to research assistants.

The skill underneath

Spaces turn a description into a tool, which makes the description the whole job. Saying exactly what you want, what you do not, and then checking the result is the same skill whether the tool is Muse, ChatGPT or Claude. That is what Coursium teaches: short lessons on giving AI clear instructions and checking its output, with practice where you write a prompt and get feedback. If you want to try that before installing anything, the free practice example runs in the browser.