ReleaseAPI / Product · ActionableTechCrunch→ original

Meta Adds “Spaces” to Muse: The Agent Builds Trackers and Planners for Each Task

Meta has launched Spaces for Muse agents: separate topic-based workspaces in the sidebar. You describe an idea in the chat, and Muse builds a tracker, planner, or project board around it and guides it toward the goal. Connected apps update the data in the background, including new concert dates, price changes, and savings progress. This is Muse’s second new feature in a week, following voice mode.

AI-processed from TechCrunch; edited by Hamidun News
Meta Adds “Spaces” to Muse: The Agent Builds Trackers and Planners for Each Task
Source: TechCrunch. Collage: Hamidun News.
Listen to article
Article Fact Sheet & Key Figures
ParameterDetails
Primary sourceTechCrunch ↗
Editor-in-Chief
Key metric #1Spaces are created from a description in the chat, with no builder required
Key metric #2App integrations update data without being prompted
Key metric #3Muse agents have 10 million users in the US and Canada

A chat with an endless stream of messages is poorly suited to long-running tasks. You spend a week planning a vacation, and three days later the link you need is already buried in the conversation. Meta is addressing precisely this problem in its Muse agents by adding Spaces.

What it is

Spaces are separate topic-based conversations that live in the sidebar of the Muse interface. Meta itself calls them “custom experiences,” but they are essentially project tabs. Open a “Wedding” or “Holiday Gifts” space, and you return to the context of a single task, where Muse works toward a goal on your behalf.

The key detail is that a space does not need to be configured manually. You describe the idea in the chat, and Muse builds the appropriate format itself: a tracker, planner, or project board. Mark Zuckerberg wrote on X that he used it to build a BBQ game, an event tracker, and a home-management dashboard.

What it looks like in practice

Suppose you are collecting recipes for a family celebration. You tell the agent: “I need a menu for 12 people, and some guests are gluten-free.” Muse creates a Space containing recipes, a shopping list, and notes. Everything is kept in one place instead of being scattered across 40 messages.

The second component is integrations. According to Meta, if you connect your apps, spaces update in the background: new concert dates appear, prices change, and savings progress increases. The agent brings you the news proactively instead of waiting for a question. In its announcement, the company did not specify exactly which apps are supported, how permissions work, or whether there is an API for third-party developers.

Why this matters to product teams and founders

First, this represents a shift in agent UX: from “chatting with a model” to a workspace with memory and state. A similar pattern is already familiar from artifacts, projects, and canvases in other assistants. Users do not need an answer; they need a living object they can return to.

Second, it is prompt-based mini-app generation. A tracker or dashboard assembled from a single sentence is essentially personal no-code. For companies making simple consumer utilities such as planners, trackers, and hobby tools, this is an uncomfortable signal: users can now get the basic version directly inside the agent.

Third, there is distribution. During an October earnings call, Zuckerberg said the agents had 10 million users in the US and Canada, less than two months after launch. That is an audience to which Meta can roll out new features without a separate app. Spaces is the second new feature in a week: one day earlier, the company announced Muse Voice, which provides video chats with the agent and voice commands.

Limitations and what to do

For now, this is an announcement of a consumer feature, not a developer platform. There is no data on the quality of the generated spaces or the reliability of background updates. Privacy also remains a concern: an agent with access to your apps creates a new surface for errors and leaks. The TechCrunch author explicitly notes that the description of Spaces is based on “as far as he can tell.”

Competitors are not standing still either: OpenAI has dots, while SpaceXAI has Grokbot, and both launched with similar cartoon characters. The battle is over whose agent will become the default entry point at home.

What to do in practice:

  • Product teams: check whether your chat is turning into an endless stream. Give users projects with persistent state.
  • Founders of consumer utilities: assess what an agent can generate from a single sentence, and shift your value toward data, integrations, and community.
  • Developers: watch whether Meta opens Spaces integrations to external services. If it does, this will be a new distribution channel.
Story

Key Questions Answered

What it is?

Spaces are separate topic-based conversations that live in the sidebar of the Muse interface. Meta itself calls them “custom experiences,” but they are essentially project tabs. Open a “Wedding” or “Holiday Gifts” space, and you return to the context of a single task, where Muse works toward a goal on your behalf. The key detail is that a space does not need to be configured manually. You describe the idea in the chat, and Muse builds the appropriate format itself: a tracker, planner, or project board.

What it looks like in practice?

Suppose you are collecting recipes for a family celebration. You tell the agent: “I need a menu for 12 people, and some guests are gluten-free.” Muse creates a Space containing recipes, a shopping list, and notes. Everything is kept in one place instead of being scattered across 40 messages. The second component is integrations. According to Meta, if you connect your apps, spaces update in the background: new concert dates appear, prices change, and savings progress increases.

Why this matters to product teams and founders?

First, this represents a shift in agent UX: from “chatting with a model” to a workspace with memory and state. A similar pattern is already familiar from artifacts, projects, and canvases in other assistants. Users do not need an answer; they need a living object they can return to. Second, it is prompt-based mini-app generation. A tracker or dashboard assembled from a single sentence is essentially personal no-code.

ZK
Zhemal Khamidun — CPO at Alpina Digital, product manager of AlpinaGPT. Building AI products since 2019. Hamidun News is a pet project that grew into a production aggregator with 6,000+ readers.
✦ Private AI Builders Community · Vibe Club

Ship AI Faster with Founders, Engineers & Weekly Architecture Reviews

Inside Vibe Club (hamidun.com/club) — weekly live masterminds with Zhemal Khamidun, Tech Radar, full LMS access, and 370+ production Claude Code skills.

Join Vibe Club →Explore Hamidun Academy
What do you think?
Loading comments…