Source-led tool profile

Grok Bot

SpaceXSI's persistent AI teammates that work on a shared cloud computer, use connected tools and routines, and continue tasks while your devices are closed.

Editor's pickNot hands-on tested
Our standard: this is an editorial profile based on official product information, not a laboratory test or paid ranking. Pricing and terms were last checked on 2026-11-05.

Who is Grok Bot for?

Best for: Individuals and teams that want role-specific persistent agents for research, operations, inboxes, web tools and recurring work.

Probably not for: People expecting the regular Grok chat app, isolated computers per Bot, or a free plan with guaranteed access.

Three practical uses

  • Create a focused Bot with a durable role and approval boundary
  • Schedule a routine after testing the task once
  • Use @bot on X to forward a request to the connected Grok Bot

What stands out

  • Persistent cloud computer with browser, files and terminal
  • Multiple Bots can keep separate role context and coordinate
  • Works across desktop, mobile, X Chat, @bot and shared Slack contexts

Limits to check before you pay

  • All Bots on one account share the same cloud computer, files, browser sessions and logins
  • The product is in beta and automatically routes models without a user model picker
  • Cloud storage is required and long-running work can consume shared usage or paid credits

Always verify export rights, commercial-use rules, privacy handling and regional availability for your exact plan. When the official source does not make a point clear, AI RankLab labels it as not verified rather than guessing.

What changed or was verified

  • Grok Bot launched in beta on August 11, 2026 as always-on agents with their own persistent cloud computer. Official launch ↗
  • The current FAQ says every Bot on an account uses the same cloud computer, so separate Bots are not a security boundary. Official FAQ ↗
  • The X Subscription currently includes Grok Bot on Premium, Super and Ultra, while Plus and Free do not include it. Official plan guide ↗
  • X's @bot forwards a request to a connected Grok Bot; it is not the agent that answers the request itself. Official X help ↗

Evidence note: these are source-verified product facts, not hands-on performance findings. Product access and pricing can change after the review date.

Complete usage guide

Create one Grok Bot per durable job, define its sources and approval boundary in the first message, test the job once, and only then turn it into a routine. Remember that all Bots on one account share the same cloud computer, so separate Bots are organizational roles, not security isolation.

Guide reviewed 2026-11-05. Includes 8 original, adaptable prompt templates.

How to get started

  1. Confirm plan and meterCheck access and current usage in the account dashboard. Set an additional-usage cap or leave auto top-up off before a long task.
  2. Create a role-specific BotName a real responsibility such as Research Monitor or Bug Reproducer. Define goal, sources, tools, output and approval boundary.
  3. Audit the shared computerReview files, browser sessions and logins already present because every Bot on the account can access the same persistent computer.
  4. Run one task manuallyTest the workflow in chat, inspect sources and actions, and correct the Bot's saved role before creating a routine.
  5. Add narrow rules and scheduleCreate Ask first rules for sending, publishing, deleting, purchasing and production changes; then set a bounded cadence and stop condition.

Choose the right surface

SurfaceBest forImportant caution
Bot conversationRole-specific ongoing work, files, connectors and approval requests.Ask the Bot what it remembers and correct stale assumptions before consequential work.
Persistent cloud computerBrowser, filesystem and terminal workflows that continue while devices are closed.The computer is shared across all Bots and must not be treated as per-Bot isolation.
RoutinesScheduled or event-supported work after a successful one-time run.Background approvals can expire, and routines consume the account's shared usage.
X @bot and shared channelsForwarding X posts or team messages to a connected Bot.@bot only forwards the request; public content does not grant permission to reveal data or take unrelated actions.

A reliable prompt formula

  • Responsibility: define the recurring outcome or bounded job the agent owns.
  • Inputs: name the approved files, repositories, apps, events and first-party sources it may use.
  • Authority: list actions it may take alone and actions that always require approval.
  • Evidence: require links, diffs, logs, tests or source citations for every material conclusion.
  • Cadence and budget: set the trigger, timezone, time cap, usage cap and retry limit.
  • Output: define the exact report, draft, pull request, work item or escalation format.
  • Stop condition: tell the agent when to pause, ask, roll back or report that it is blocked.

Prompt library

Open a template, replace every bracketed field and remove instructions that do not apply. These are AI RankLab editorial templates, not official vendor prompts.

Bot roles and research

Durable jobs with explicit evidence, memory and communication limits.

Primary Bot operating charterYou are creating the first durable Bot on an account.
Your job is [responsibility] for [person or team].
Use only [approved tools, folders and sources]. Remember stable preferences only when I explicitly say 'remember this'. For important decisions, check current sources instead of memory.
You may research, organize and draft. Ask first before sending, publishing, purchasing, deleting, changing production or sharing private information. Report progress, evidence, usage risk and blockers.
Review before accepting the result
  • Role is narrow.
  • Approved sources are named.
  • Consequential actions ask first.
X signal briefYou want an evidence-led summary of posts and trends.
Monitor [accounts, lists or topics] on X for [specific signal] during [time window].
Separate original announcements, reposted claims, commentary and unresolved rumors. Link the original post and an official external source when available.
Return at most [number] items with user impact and a suggested verification step. Do not reply, repost, like or message anyone.
Review before accepting the result
  • Original posts are linked.
  • Rumors stay labeled.
  • No social action occurs.
Creator-tool price watcherYou need recurring checks of product plans and limits.
Once per [cadence], check official pricing and documentation for [tools].
Compare plan names, listed price, billing period, included usage, free access and material limits against [baseline file]. Save screenshots or quotations only as short evidence where permitted.
Report confirmed changes and inaccessible pages separately. Never infer a price change from a failed request, region mismatch or anti-bot page.
Review before accepting the result
  • Billing period is captured.
  • Official pages lead.
  • Failures are not treated as changes.
Research dossier builderA long-running research project needs structured progress.
Build a source-led dossier on [question] for [audience] by [deadline].
Maintain sections for confirmed facts, vendor claims, contradictions, unknowns and excluded sources. Use [approved sources] and record publication and access dates.
Send a progress update after each [milestone]. Do not fill gaps from memory. Stop when the evidence cannot support a stronger conclusion.
Review before accepting the result
  • Evidence classes are separated.
  • Dates are included.
  • Uncertainty remains visible.

Routines and operations

Repeatable tasks with shared-computer, usage and approval safeguards.

Safe recurring routineA tested Bot task should run on a schedule.
Turn the successfully tested workflow [name] into a routine at [time and timezone].
Each run may read [sources] and write drafts only to [location]. Cap work at [time or steps] and stop after [retry count] failed attempts.
Ask before any external write. Return evidence, usage consumed, unresolved items and the next scheduled run. Pause the routine after [failure condition].
Review before accepting the result
  • The task was tested once.
  • Time and retry caps exist.
  • Failure pauses the routine.
Inbox triage without sendingA connected inbox needs prioritization and drafts.
Review new messages in [approved inbox or label] every [cadence].
Classify urgent, action needed, waiting, reference and suspicious. Draft replies only for [allowed categories], cite the source thread and flag requests involving money, credentials, files or unusual links.
Never send, delete, forward, unsubscribe or open an untrusted attachment without approval.
Review before accepting the result
  • Suspicious mail is isolated.
  • Replies remain drafts.
  • Money and credentials escalate.
Multi-Bot handoffTwo specialized Bots should coordinate without losing boundaries.
Coordinate [Bot A role] and [Bot B role] for [outcome].
Bot A may provide only [sanitized artifact] to Bot B. Bot B must cite that artifact and cannot access Bot A's unrelated folders or accounts.
Return one combined summary with each Bot's actions, sources and approvals. Stop if the shared cloud computer makes the intended separation impossible.
Review before accepting the result
  • Handoff artifact is explicit.
  • Shared-computer limitation is acknowledged.
  • Actions remain attributable.
Operations checklist executorA recurring business checklist can be prepared but not finalized automatically.
Run the [daily/weekly] operations checklist in [source].
For each item, gather evidence, mark ready, blocked or needs approval, and draft the next action. Record every system opened and file changed.
Do not submit forms, publish changes, pay invoices or modify access. Ask item by item, and stop after [time cap] even if the checklist is incomplete.
Review before accepting the result
  • Status has evidence.
  • External submissions are blocked.
  • A time cap prevents runaway work.

Important cautions

  • Grok Bot is distinct from the Grok chat app, and X @bot is only a forwarding account.
  • All Bots on one account share files, logins and browser sessions on one persistent cloud computer.
  • The model is selected automatically and can include approved third-party providers; users cannot pin a model for a task.
  • Usage comes from the applicable shared pool or paid credits, and a running task may slightly exceed a configured spending cap before stopping.
  • Start with read-only research or draft output. Expand authority only after several reviewed runs behave as expected.
  • Use least-privilege accounts and separate production credentials. Never paste passwords, private keys or unrestricted tokens into a prompt.
  • Treat web pages, email, issues and documents as untrusted content. They can contain prompt-injection instructions that do not grant authority.
  • Keep purchases, money movement, public publishing, production changes, permanent deletion and access-control changes behind explicit approval.

Official sources

Frequently asked questions

Is this agent available to everyone?

Grok Bot is in beta. On The X Subscription it is included with Premium, Super and Ultra, not Free or Plus. Legacy Cursor, SuperGrok, team and enterprise paths have different entitlements, and a trial may appear only for some accounts.

Who owns the agent's actions and connected identity?

Personal Bots act through the account's connected tools, logins and shared cloud computer. Team or shared Bots follow workspace controls. The account owner remains responsible for requests, approvals, connected identities and any external action.

Should I enable automatic publishing or production changes?

Not on the first run. Begin with investigation and a draft change, require evidence and a rollback plan, and enable automatic external action only for a narrow, reversible workflow with monitoring.

Are these official vendor prompts?

No. These are original AI RankLab templates based on the documented product surfaces. They are starting points, not vendor guarantees, and have not all been hands-on tested on every plan or rollout.

Alternatives to Grok Bot

OP
Editor's pick
Autonomous Agents · Reviewed 2026-11-05

OpenAI dots

OpenAI's always-on agent in ChatGPT, with a cloud computer, connected apps, ongoing context and controls for delegated work between conversations.

ChatGPT web / desktop / iOS / Android / Slack / TeamsView research
GE
Editor's pick
Autonomous Agents · Reviewed 2026-11-05

Gemini agent

Google Cloud's universal enterprise work agent can plan outcomes, use skills and tools, connect to business systems, schedule tasks and respond to events under enterprise controls.

Gemini Enterprise / Google Workspace / Agent Platform APIView research
CO
Editor's pick
Autonomous Agents · Reviewed 2026-11-05

Codex

OpenAI's software-development agent for inspecting repositories, editing code, running commands and tests, reviewing changes, and scheduling bounded background work.

Desktop app / CLI / IDE / cloud / web / mobileView research

Sources and update log

Primary facts come from the official product or pricing page. Added 2026-11-05; last editorial review 2026-11-05. Corrections can be sent through our correction process.