# Grok Bot (slim)

> Hosted assistant from SpaceXAI and Cursor. Named Bots work on a persistent cloud computer with a browser, files and terminal, reached from desktop and mobile apps, Slack and X, with routines that run on a schedule.

- Full: https://www.anchorterminal.com/tools/grok-bot.md (~10,400 tokens) · this version ~2,080 tokens · JSON https://www.anchorterminal.com/tools/grok-bot.json · canonical https://www.anchorterminal.com/tools/grok-bot
- Index: https://www.anchorterminal.com/llms.txt · API: https://www.anchorterminal.com/api/v1/index.json · Updated: 2026-10-12

**E · 45.2/100 · rank #893 of 986 · #12 in Personal assistants · not agent-ready · confidence medium**


Assessment: Connector tokens stay on Cursor's backend and each user gets a separate microVM, with approvals, drafts and a review model before risky actions. All of a user's Bots share one computer and its sign-ins, the review model skips memory writes, and the status page shows seven major Grok Bot incidents since 3 September 2026.

## Facts

- Kind: Platform · vendor: SpaceXAI and Cursor (Anysphere, Inc.) · category: Personal assistants · legal entity: Anysphere, Inc. · provenance 84/100
- Auth: OAuth or key · pricing: Paid · x402: no · licence: Proprietary
- Probe metrics: not measured yet (probes haven't run)
- What it does without asking: Runs routines on a schedule or after an event, works in the background on the cloud computer, starts a task from an @bot tag on X, and lets Bots message each other and pass tasks on. Which actions stop for approval depends on the tool, the risk and the owner's Auto-review rules. No fixed list of always-gated actions was found (https://docs.x.ai/grok-bot/faq)
- Approval controls: Allow once, Always allow and Deny on each request. Personal Ask first and Allow automatically rules, with Ask first winning. Editable drafts for email and Slack messages. Unattended approvals expire after about 10 minutes. Local commands on the owner's own machine ask every time by default. Enterprise admins can enforce Auto-review and add team rules (https://docs.x.ai/grok-bot/approvals-security-and-privacy)
- Accounts and credentials: Connectors are installed as plugins and authenticated in the browser. OAuth tokens stay on Cursor's backend and are not given to the Bot or stored on the computer. Sites without a connector are used through the browser, where the owner signs in by taking control, and those sessions persist on the shared computer. Connector scopes are not listed on the pages read (https://docs.x.ai/grok-bot/security; https://docs.x.ai/grok-bot/computer-and-apps)
- Memory: A Bot keeps stable preferences, role context and summaries of prior work, separate for each Bot. A duplicate does not copy learned memory. Team Bots keep team memory plus private notes for each person. Auto Review does not check memory writes. No screen to view or delete single memories is described on the pages read (https://docs.x.ai/grok-bot/bots; https://docs.x.ai/grok-bot/team-bots)
- Deletion: Deleting a Bot removes its profile, conversation and routines, not the shared files or browser sessions. Account deletion removes agents, chats, computers and plugins within 30 days. Under the DPA, data is deleted or returned within 30 days of written direction after the service ends. No retention policy per organisation (https://docs.x.ai/grok-bot/security; https://cursor.com/help/grok-bot/delete-account)
- Where it runs: One Firecracker microVM per user in Cursor's cloud, in the United States today, with no written residency commitment by default. Cursor picks the model. Its status page names Grok, Anthropic and OpenAI models among those it serves, and the docs guarantee no fixed vendor set (https://docs.x.ai/grok-bot/security; https://status.cursor.com/history)
- Channels: Desktop app for macOS, Windows and Linux, mobile app for iPhone, iPad and Android, with text, dictation, voice chat and voice memos. Tagging @bot on X. Slack for Team Bots. No SMS, WhatsApp, Telegram or email channel is described on the pages read (https://docs.x.ai/grok-bot/faq)
- Cost: No separate subscription. Included with paid individual Cursor plans (the pricing page shows Individual at $20 a month) and Cursor Teams ($40 a user a month), or through a linked SuperGrok, SuperGrok Plus, SuperGrok Heavy or X Premium+ subscription. Weekly included usage with no published amount, then on-demand usage billed through Cursor (https://cursor.com/help/grok-bot/plans; https://cursor.com/pricing)
- MCP client: Connectors follow the team's Cursor MCP policy. Team Bots accept custom MCP servers over remote HTTPS or as a command run on the conversation's computer (https://docs.x.ai/grok-bot/team-bots)
- Logging: The transcript shows tool activity, computer use, files and approval requests. Audit logs, Action Recording (off by default, 90-day retention) and OpenTelemetry export are Enterprise only (https://docs.x.ai/grok-bot/security)
- Status: status.cursor.com, with a Grok Bot component. Sixteen incidents named Grok Bot from 18 August to 9 October 2026, seven marked major (checked 2026-10-12)
- Prices: Cursor Individual plan, as shown on the pricing page $20 per month (plan); Cursor Teams plan $40 per seat per month

- Scores: Reliability 40, Performance pending, Schema & documentation 40, Agent ergonomics 50, Security & auth 68, Payments & pricing 25, Task success pending, Maintenance & community 20, Transparency & trust 66 · total over the 7 assessed categories
- Why: Reliability, Read with the hosted lines. · Schema & documentation, Grok Bot is an app, so the API lines are read for what a person or an agent setting it up can rely on. · Agent ergonomics, Read for an owner directing the assistant. · Security & auth, Read with the tool lines, applied to the accounts the assistant holds. · Payments & pricing, No machine payment protocol (0). · Maintenance & community, No dated release notes or changelog for Grok Bot were found. · Transparency & trust, Closed source with clear terms.
- Sources: 31, open questions: 15, both in the full twin
- Capabilities: assistant.proactive, assistant.accounts, assistant.memory, assistant.messaging, agent.mcp-client
- JSON: https://www.anchorterminal.com/api/v1/tools/grok-bot.json
- Verify (for the vendor): the badge `https://www.anchorterminal.com/badges/grok-bot.svg` or a link to https://www.anchorterminal.com/tools/grok-bot from a page on cursor.com or x.ai or one of their subdomains, then `POST https://www.anchorterminal.com/api/v1/verify` `{"slug", "url"}` or `verify_listing` at /mcp; re-checked weekly, no effect on the grade. Snippets in the full twin.

## Before you call it

1. State in each request what the Bot may change and where it must stop. Which actions pause for approval depends on the tool, the risk and the owner's Auto-review rules
2. Add narrow Ask first rules for sending, publishing, purchasing and deleting. Ask first wins over Allow automatically when both match
3. Treat any sign-in or file on the cloud computer as available to every Bot on the account. Give a workload its own Cursor user when it needs separate credentials
4. Run Test run with safe inputs. A test run performs real work and can call connected tools
5. To remove access, pause routines, sign out of sites on the computer, uninstall connectors and revoke them at the source service. Deleting the Bot is not enough

## Similar tools

| Tool | Grade | Score | Shared capabilities | Slim |
| --- | --- | --- | --- | --- |
| ZeroClaw | BB | 70.2 | assistant.proactive, assistant.messaging, assistant.memory, assistant.accounts, agent.mcp-client | https://www.anchorterminal.com/tools/zeroclaw.min.md |
| NanoClaw | B | 68.3 | assistant.proactive, assistant.messaging, assistant.memory, agent.mcp-client, assistant.accounts | https://www.anchorterminal.com/tools/nanoclaw.min.md |
| Hermes Agent | B | 67.9 | assistant.proactive, assistant.messaging, assistant.accounts, assistant.memory, agent.mcp-client | https://www.anchorterminal.com/tools/hermes-agent.min.md |
| OpenClaw | B | 67.1 | assistant.proactive, assistant.messaging, assistant.accounts, assistant.memory, agent.mcp-client | https://www.anchorterminal.com/tools/openclaw.min.md |
| IronClaw | B | 66.7 | assistant.proactive, assistant.messaging, assistant.accounts, assistant.memory, agent.mcp-client | https://www.anchorterminal.com/tools/ironclaw.min.md |


## Panel reviews (0, desk reviews from public material, no calls made)