Grok Bot agent gets its own @mail.grokbot.com email address, rolling out to users now

Grok Bot now has its own email address. The agent can use an inbox ending in @mail.grokbot.com to sign up for services, contact businesses on a user's behalf and schedule time with other people, according to the announcement on X. The feature began rolling out to users on October 9, 2026, per 9to5Mac.

How to claim the inbox

The announcement says users can claim an address by asking Bot directly or by tagging @bot on X. Prompting Grok Bot with something like "claim your email address" starts the process in the chat interface.

Users can request a custom address, and Grok Bot will check whether it is available. The address is used by the agent system, meaning it is the agent's own identity, not a forwarding alias for the user's personal mailbox.

What the agent can do with it

The announcement lists three use cases:

• Signing up for services
• Contacting businesses for the user
• Scheduling time with someone

The announcement does not say how much autonomy the agent has when sending mail. It also does not say whether users must approve outgoing messages, or what retention or privacy rules apply to the inbox. None of these details have been disclosed.

Context: second upgrade this week

The email feature is Grok Bot's second significant upgrade this week. 9to5Mac reports that earlier in the week the agent received an intelligence upgrade "with the help of Anthropic's Claude." The report does not specify which Claude model is involved or how it is integrated.

Grok Bot is available on iPhone, iPad and Mac. Pricing, tier availability and regional restrictions for the email feature have not been disclosed.

Competitive backdrop

9to5Mac places the update alongside two rival agent efforts:

• Meta is pushing its mostly free Muse agent, which is newly available on iPad.
• OpenAI is building ChatGPT Dot, described as an Astra-powered proactive agent.

These details come from 9to5Mac's reporting, and the source does not provide further specifications for either product.

What this means

Giving an agent a persistent email identity is a practical step beyond chat-based assistance. Email is the default credential and communication layer for most online services. An agent with its own address can register for accounts and hold ongoing threads with third parties without touching the user's personal inbox. That separation also limits exposure if the agent mishandles a message.

The open questions are about control and safety. Autonomous outbound email raises issues around spam filtering, impersonation, prompt injection through inbound messages, and whether recipients can tell they are talking to an agent. Until the developer documents approval flows and data handling, the feature's reliability and risk profile remain unverified.

The update also shows the agent race moving from model quality toward capabilities such as identity, communication and scheduling.