Meta gives Muse agents email addresses and video chat in rapid update push
Meta's Muse AI agents can now send and receive email and hold video conversations, as the company races to expand what its newest assistant can do.
Meta is expanding its Muse AI agent less than a month after introducing it, rolling out a set of updates that give the assistant its own email address and the ability to talk with users over video. The changes, announced by the company today, push Muse from a text-based helper toward something closer to a digital coworker that can act on a user's behalf across apps and services.
The email feature is the more consequential of the two. Each Muse agent gets a dedicated address, which means it can receive messages, reply to them, and use correspondence as an input for completing tasks. That turns the agent into a participant in workflows rather than a tool you have to feed instructions to manually. Meta has not said exactly which email providers or clients are supported, or how it handles spam, permissions, and account recovery for an address that belongs to software rather than a person.
Video chat is the second addition, letting users speak with their Muse face to face instead of typing prompts. Meta has been building toward more natural interaction across its products, and a camera-based conversation is a logical step for an assistant that already handles voice and text. The company did not detail latency, supported devices, or whether the video stream is processed on-device or in the cloud, details that will matter to users with privacy concerns.
The timing is aggressive. Meta unveiled Muse only recently, and shipping two notable capabilities this quickly suggests the company wants to establish a habit of use before rivals lock in their own agent ecosystems. OpenAI, Google, and Anthropic have all been adding agentic features — tool use, memory, and the ability to take actions — and email access is one of the clearest ways to demonstrate an assistant that does work instead of just describing it.
Email also raises the stakes on security. An agent that can read and send mail on your behalf is a target for phishing, prompt injection, and impersonation, and a compromised address could be used to reach the user's contacts. Meta will need to explain how it separates instructions from untrusted content and what a user can do if the agent sends something it should not have. Those answers are not yet public.
What to watch next is whether Meta opens Muse's email capability to third-party services and how it prices or tiers the video feature. The company has a long history of shipping fast and refining later, so expect more updates soon — and expect regulators and security researchers to look closely at an assistant that now has a mailbox of its own.