📅 November 6, 2026 • TechCrunch
Meta adds Spaces to its Muse agent
Meta introduced Spaces, a new way to collaborate with the Muse agent, per TechCrunch. Spaces are custom experiences inside the Muse sidebar - trackers, planners and project boards that the agent assembles from a plain chat description. The flow per Meta - describe an idea in chat, and Muse builds a space for you.
From chat description to mini-app
The core mechanic is natural language acting as an app builder. A user describes what needs tracking - a training plan, a project, a budget - and Muse materializes the corresponding space in the sidebar. No template selection and no form filling - the structure is generated from the conversation itself, and the result behaves like a small application rather than a chat thread. Spaces live in the Muse agent sidebar, so a tracker assembled from one chat description stays in place as a standing object rather than a one-off answer.
Third-party data flows in the background
Spaces connect to third-party applications, and updates arrive without a user request - new concert dates, price changes and savings progress are pulled into the space in the background. The space works as a live surface instead of a static page - the agent watches external sources and refreshes the content on its own.
Adoption scale and release cadence
Per Meta, Muse agents reached 10 million users in the US and Canada in under two months after launch - the figure comes from Zuckerberg on the October earnings call. Spaces is the second new Muse feature this week - on November 5 Meta announced Muse Voice, which adds video chat with the agent and voice commands. Adoption figures reflect early usage per Meta and do not describe future capabilities.
IMPACT
For practitioners building user-facing agent products, Spaces documents a pattern worth tracking - the interface layer of an agent is shifting from a chat transcript toward persistent workspaces that the agent itself constructs and maintains. Daily work changes first for end users - routine tracking moves from generic chat threads into purpose-built surfaces fed by external data. Teams designing agent features get a concrete example of two ideas working together - generation of mini-apps from natural language and background synchronization with third-party services. The 10 million user figure per Meta sets the adoption baseline for this product category. The background data pull also changes what an agent session means - updates arrive while the user is absent, so a space reflects the current state of external services rather than the state at the last prompt.
Sources: techcrunch.com
Continue the conversation: https://t.me/AIAutomationArchitecture · Bluesky
Top comments (0)