什么是 Grok Bot

Bots are AI teammates you can give real work to. They can sign into and use apps and websites just like you do, on a persistent cloud computer. They can collaborate independently, passing context between each other and handing off tasks, and they understand the nuances of how you like work done over time. They finish jobs end to end, and only come back when something needs your approval. You work with a Bot by messaging it like a teammate: give it a task or function, relevant context, and access to any tools or files it needs. Bots take on ambitious projects, work across multiple tools and systems, and keep you updated of their actions in the conversation. In the Docs and in the Grok Bot app, a Bot is a single persistent, named agent—one AI teammate.

Bot 是能交真活的 AI 同事。它们有一台一直开着的云电脑，能登录、使用你平时用的 App 和网站。它们能自己协作，互相传上下文、交接任务；干久了，也懂你喜欢怎么干。活是从头干到尾的，只有需要你拍板时才回来。你跟它说话就像给同事发消息：说清任务、相关背景，再给它需要的工具或文件。它可以接大项目，跨很多系统干活，并在对话里随时告诉你它做了什么。文档和 App 里，一个 Bot = 一个持久、有名字的 agent，也就是一位 AI 同事。

The 11 August 2026 launch post calls Grok Bot “your team of always-on agents.” They have their own computer, work inside tools and apps like you do, and keep working 24/7. You can message them like a colleague and hand off the work. They remember conversations, learn how you like things done, and get sharper the more you work together. SpaceXAI first built it as an internal prototype; teams created Bots for sales outbound, marketing campaigns, office operations, bug fixes, and more. Product teammate Roman put the difference this way: “There is a huge difference between 90% done and 100% done. Most AI gets you almost there. Grok Bot can finish the swing, because the work lands where a human would put it, in the actual tool.” Community teammate Fiona said there was nothing to learn: “It was just like bringing on a coworker.”

2026 年 8 月 11 日的介绍文把 Grok Bot 叫「一直在线的 agent 团队」。它们有自己的电脑，进你已经在用的工具里干活，全天不停。你可以像给同事发消息一样交活。它们记得对话，会学你喜欢的做法，一起干得越多越顺手。这东西先是 SpaceXAI 内部原型：做出站销售、营销活动、办公室事务、修 bug，都有人拿 Bot 干。产品同事 Roman 说：「做到 90% 和做到 100% 差很多。多数 AI 只送到门口。Grok Bot 能把这一棒打完，因为活落在真人会放的地方——真实工具里。」社区同事 Fiona 说没什么要学的：「就像新来一个同事。」

Each Bot runs on a persistent cloud VM with a browser, filesystem, and terminal. It can use connectors or MCP where available, and computer use for apps and websites without a clean API, so work finishes in the real tools rather than as chat drafts. Getting started is: create a Bot, message it, and grant access as needed. No workflow builder or prior Bot setup is required. The same Bot is reachable from the desktop app and iOS.

每个 Bot 跑在一台持久的云虚拟机上，自带浏览器、文件系统和终端。有连接器或 MCP 就用；没有干净 API 的 App 和网站，就用电脑操作去点。活是在真实工具里收尾，不是只给你一份聊天草稿。上手就是：建一个 Bot，给它发消息，需要时再授权。不用先搭工作流，也不用预先配置。同一个 Bot，桌面 App 和 iPhone 都能找到。

Multiple Bots share one user-scoped computer and can run in parallel. They can message each other, share context in threads or group chats, and pass ownership so you are not the router between tools. You can also ask a Bot to follow along once through a multi-step or multi-system path; the overview says it persists that path as a routine and can re-run it on a schedule or on demand. Named Bots keep memory, files, browser sessions, and preferences across turns. Context compounds instead of resetting to a fresh environment on every task.

多个 Bot 共用你账号下的一台电脑，可以并行干活。它们能互相发消息，在话题或群聊里共享上下文，并把所有权交出去，这样你就不用在工具之间当路由器。你也可以让 Bot 跟着看一遍跨步骤、跨系统的流程；概览写它会把这条路存成 routine，之后能按日程或按你的指令再跑。有名字的 Bot 会跨轮次保住记忆、文件、浏览器登录和偏好。上下文是越积越多，不是每接一个活就清空环境。

All of your Bots use the same persistent cloud computer. They share files, browser sessions, and app logins, which makes handoffs possible without repeating setup. The computer is isolated to your account, not to an individual Bot. Treat a login or file placed on the computer as available to all of your Bots. Each Bot gets its own screen on that computer, so several Bots can use browser and desktop tools in parallel—but those screens are not separate security boundaries.

你名下所有 Bot 用同一台持久云电脑。文件、浏览器会话、App 登录都是共享的，所以交接时不用重新配置。这台电脑隔离在你的账号，不是某个 Bot 独占。你放上去的登录或文件，就当所有 Bot 都能用。每个 Bot 在这台电脑上有自己的屏幕，所以几个 Bot 可以同时用浏览器和桌面工具；但屏幕分开，不等于安全边界分开。

Start with a real task that involves multiple tools and has a clear result. The docs example: pull this week's Strategic Prospects PG List from Salesforce; skip anyone already in a sequence; research the top 5 accounts across the web, Slack, Databricks, and Sumble; pull contacts; draft LinkedIn and email in your voice; and leave drafts to approve by tomorrow morning. That tells the Bot what to do, where to work, what context to pull in, and what the finished deliverable should look like. You can review the work, provide corrections, and later turn the process into a reusable skill or a proactive routine.

第一次就交一个真任务：要动好几个工具，结果也要清楚。文档的例子是：从 Salesforce 拉本周 Strategic Prospects PG 名单，已经在 sequence 里的跳过；在网上、Slack、Databricks、Sumble 研究前 5 个客户，找联系人，用你的口吻起草 LinkedIn 和邮件，明天早上留草稿给你批。这等于告诉它：干什么、在哪干、要带上什么背景、成品长什么样。你看完可以改，之后还能把流程存成可复用的 skill，或做成会主动跑的 routine。