Grok Bot comparison
Grok Bot vs Cloud Agents: Which Should You Use?
Use Grok Bot for persistent cross-app work owned by a named teammate, such as research, reconciliation, and recurring preparation. Use a Cursor Cloud Agent for a bounded repository task that should build, test, and return a branch or pull request. Grok Bot can coordinate the wider outcome and launch a Cloud Agent for its coding step.
On this page
- What is the main difference between Grok Bot and Cursor Cloud Agents?
- How do their cloud workspaces and persistence differ?
- What does each product return for review?
- Which has stronger security and governance controls?
- How do privacy, model choice, and billing compare?
- When should you choose Grok Bot or Cursor Cloud Agents?
- How can Grok Bot and Cursor Cloud Agents work together?
What is the main difference between Grok Bot and Cursor Cloud Agents?
Grok BotAn early AI teammate product from SpaceXAI and Cursor that works across apps on a persistent cloud computer and asks for approval on selected actions; launch materials describe it as beta, but Cursor says Beta Services terms do not apply. Press Enter for the full definition. owns a standing job across tools. A Cursor Cloud Agent owns one repository task in a development environment. Both work in the cloud and can keep running without your laptop, but they start from different units of work and return different kinds of evidence.
- Dimension
- Primary unit
- Grok Bot
- A named teammate with a durable job and conversation
- Cursor Cloud Agent
- A bounded coding task
- Dimension
- Workspace
- Grok Bot
- One persistent user-scoped computer shared by all of that user's Bots
- Cursor Cloud Agent
- One isolated VM for the agent run, with configured repositories and dependencies
- Dimension
- Typical systems
- Grok Bot
- Business apps, websites, files, terminal, connectors, and MCPModel Context Protocol. A standard that lets an AI agent pull in context from outside the repo, like Jira tickets or internal docs. Press Enter for the full definition.
- Cursor Cloud Agent
- Git repositories, test tools, services, browser, terminal, MCPModel Context Protocol. A standard that lets an AI agent pull in context from outside the repo, like Jira tickets or internal docs. Press Enter for the full definition., and development dependencies
- Dimension
- Typical output
- Grok Bot
- A report, spreadsheet, draft, action, or recurring result
- Cursor Cloud Agent
- A branch, draft pull request, code diff, test output, and artifacts
- Dimension
- Persistence
- Grok Bot
- Named role, memory, files, browser sessions, and preferences across work
- Cursor Cloud Agent
- Conversation and artifacts persist; runtime resources follow Cloud Agent lifecycle and snapshot rules
- Dimension
- Model choice
- Grok Bot
- Managed by the product with automatic failover
- Cursor Cloud Agent
- User selects from a curated model set and supported context sizes
| Dimension | Grok Bot | Cursor Cloud Agent |
|---|---|---|
| Primary unit | A named teammate with a durable job and conversation | A bounded coding task |
| Workspace | One persistent user-scoped computer shared by all of that user's Bots | One isolated VM for the agent run, with configured repositories and dependencies |
| Typical systems | Business apps, websites, files, terminal, connectors, and MCPModel Context Protocol. A standard that lets an AI agent pull in context from outside the repo, like Jira tickets or internal docs. Press Enter for the full definition. | Git repositories, test tools, services, browser, terminal, MCPModel Context Protocol. A standard that lets an AI agent pull in context from outside the repo, like Jira tickets or internal docs. Press Enter for the full definition., and development dependencies |
| Typical output | A report, spreadsheet, draft, action, or recurring result | A branch, draft pull request, code diff, test output, and artifacts |
| Persistence | Named role, memory, files, browser sessions, and preferences across work | Conversation and artifacts persist; runtime resources follow Cloud Agent lifecycle and snapshot rules |
| Model choice | Managed by the product with automatic failover | User selects from a curated model set and supported context sizes |
The simplest decision is whether the job is a standing cross-tool outcome or a bounded repository change.
The overlap is real. Cloud AgentsAgents that run in a Cursor-managed virtual machine, check out the repo, do the work and open a pull request, then shut down, with no load on your laptop. Press Enter for the full definition. have a browser, computer use, MCPModel Context Protocol. A standard that lets an AI agent pull in context from outside the repo, like Jira tickets or internal docs. Press Enter for the full definition. servers, and multi-repo environments. Grok BotAn early AI teammate product from SpaceXAI and Cursor that works across apps on a persistent cloud computer and asks for approval on selected actions; launch materials describe it as beta, but Cursor says Beta Services terms do not apply. Press Enter for the full definition. has a terminal and can work with files. Capability lists alone therefore produce a muddy comparison. Look at the durable owner and the handoff. If the job should end in a reviewable code change, start with a Cloud Agent. If the job should recur across systems under one named role, start with Grok Bot.
Grok BotAn early AI teammate product from SpaceXAI and Cursor that works across apps on a persistent cloud computer and asks for approval on selected actions; launch materials describe it as beta, but Cursor says Beta Services terms do not apply. Press Enter for the full definition. can also launch a Cursor cloud agent when the team's Cloud AgentsAgents that run in a Cursor-managed virtual machine, check out the repo, do the work and open a pull request, then shut down, with no load on your laptop. Press Enter for the full definition. setting permits it. That is the clean combination: the Bot owns the wider workflow and delegates the repository change to the surface built to clone, test, and push code.
This exact topic is a hands-on Lesson: Cloud Agent Basics — about 6 minutes, free to read.
How do their cloud workspaces and persistence differ?
Grok BotAn early AI teammate product from SpaceXAI and Cursor that works across apps on a persistent cloud computer and asks for approval on selected actions; launch materials describe it as beta, but Cursor says Beta Services terms do not apply. Press Enter for the full definition. assigns one managed Linux computer to each user. Every Bot for that user shares its /workspace, browser sessions, command-line credentials, and local-computer permissions, though each Bot has its own screen. The computer is part of the long-lived teammate experience, so one Bot can continue from files or sessions another left behind.
A Cloud Agent gets an isolated development VM for its run. Cursor clones the authorized repositories, installs the configured dependencies, exposes approved secrets and network access, and lets the agent build and test there. One Cloud Agent cannot see another agent's code, environment, or state. That per-agent isolation is a direct contrast with Grok BotAn early AI teammate product from SpaceXAI and Cursor that works across apps on a persistent cloud computer and asks for approval on selected actions; launch materials describe it as beta, but Cursor says Beta Services terms do not apply. Press Enter for the full definition.'s deliberately shared user computer.
Cloud Agent persistence is split by data type. Runtime resources hibernate and are deleted on lifecycle timers, while encrypted snapshots can remain for a rolling 90 days of inactivity. Conversation state and artifacts are kept by default so a run can be reviewed or resumed, subject to deletion and enterprise retention controls. Grok BotAn early AI teammate product from SpaceXAI and Cursor that works across apps on a persistent cloud computer and asks for approval on selected actions; launch materials describe it as beta, but Cursor says Beta Services terms do not apply. Press Enter for the full definition.'s docs describe durable state but defer backend retention to the applicable Cursor terms rather than publishing a separate product-specific window.
Two Grok Bots with different jobs still share the same user's files and sign-ins. Two Cursor Cloud Agent runs use separate VM boundaries. Choose source-system scopes accordingly.
What does each product return for review?
A Cloud Agent is designed to hand code back through source control. It works on a separate branch, pushes changes, and can open a draft pull request. Screenshots, videos, logs, transcripts, run events, environment details, and setup logs can support review. Agent commits are signed and visible in Git history, and team members still need repository access to open a shared run.
Grok BotAn early AI teammate product from SpaceXAI and Cursor that works across apps on a persistent cloud computer and asks for approval on selected actions; launch materials describe it as beta, but Cursor says Beta Services terms do not apply. Press Enter for the full definition. returns work in its conversation and shared computer. That may be a document, spreadsheet, slide deck, folder of screenshots and logs, draft message, source-linked brief, or an action waiting for approval. The review contract comes from the request. If you need facts separated from inferences or a source link for every claim, say so before the run.
- Desired handoff
- A tested code change on a branch
- Better starting surface
- Cursor Cloud Agent
- Why
- Repository isolation, development environment, tests, artifacts, and draft PR handoff are native
- Desired handoff
- A weekly source-linked account review
- Better starting surface
- Grok BotAn early AI teammate product from SpaceXAI and Cursor that works across apps on a persistent cloud computer and asks for approval on selected actions; launch materials describe it as beta, but Cursor says Beta Services terms do not apply. Press Enter for the full definition.
- Why
- The named role and recurring cross-app context persist
- Desired handoff
- A bug reproduction pack with no code change
- Better starting surface
- Grok BotAn early AI teammate product from SpaceXAI and Cursor that works across apps on a persistent cloud computer and asks for approval on selected actions; launch materials describe it as beta, but Cursor says Beta Services terms do not apply. Press Enter for the full definition.
- Why
- It can inspect the issue, staging site, browser, and evidence sources as one standing job
- Desired handoff
- A bug fix after reproduction
- Better starting surface
- Cursor Cloud Agent
- Why
- The task now centers on repository changes and verification
- Desired handoff
- An operations workflow that includes one code change
- Better starting surface
- Both
- Why
- Let Grok BotAn early AI teammate product from SpaceXAI and Cursor that works across apps on a persistent cloud computer and asks for approval on selected actions; launch materials describe it as beta, but Cursor says Beta Services terms do not apply. Press Enter for the full definition. coordinate and send the code step to a Cloud Agent
| Desired handoff | Better starting surface | Why |
|---|---|---|
| A tested code change on a branch | Cursor Cloud Agent | Repository isolation, development environment, tests, artifacts, and draft PR handoff are native |
| A weekly source-linked account review | Grok BotAn early AI teammate product from SpaceXAI and Cursor that works across apps on a persistent cloud computer and asks for approval on selected actions; launch materials describe it as beta, but Cursor says Beta Services terms do not apply. Press Enter for the full definition. | The named role and recurring cross-app context persist |
| A bug reproduction pack with no code change | Grok BotAn early AI teammate product from SpaceXAI and Cursor that works across apps on a persistent cloud computer and asks for approval on selected actions; launch materials describe it as beta, but Cursor says Beta Services terms do not apply. Press Enter for the full definition. | It can inspect the issue, staging site, browser, and evidence sources as one standing job |
| A bug fix after reproduction | Cursor Cloud Agent | The task now centers on repository changes and verification |
| An operations workflow that includes one code change | Both | Let Grok BotAn early AI teammate product from SpaceXAI and Cursor that works across apps on a persistent cloud computer and asks for approval on selected actions; launch materials describe it as beta, but Cursor says Beta Services terms do not apply. Press Enter for the full definition. coordinate and send the code step to a Cloud Agent |
One workflow can cross the boundary. Keep each handoff explicit so ownership remains visible.
The review surface matters more than the flashiest capability. Git gives a code change a mature diff, branch policy, and merge gate. A spreadsheet or CRM reconciliation needs source links, totals, and an action log instead. Asking one product to imitate the other's review system usually leaves you with weaker evidence.
Which has stronger security and governance controls?
They have different control planes. Cloud AgentsAgents that run in a Cursor-managed virtual machine, check out the repo, do the work and open a pull request, then shut down, with no load on your laptop. Press Enter for the full definition. have the more complete published controls for repository-scoped engineering: per-agent VM isolation, repository authorization, network egress modes, runtime-secret redaction, hooks, signed commits, run diagnostics, retention policies, and draft pull-request handoff. Grok BotAn early AI teammate product from SpaceXAI and Cursor that works across apps on a persistent cloud computer and asks for approval on selected actions; launch materials describe it as beta, but Cursor says Beta Services terms do not apply. Press Enter for the full definition. emphasizes human credential takeover, per-action approval, Auto Review, MCPModel Context Protocol. A standard that lets an AI agent pull in context from outside the repo, like Jira tickets or internal docs. Press Enter for the full definition. policy, and a shared computer that can work across many business systems.
- Control question
- Can one run see another run's workspace?
- Grok Bot
- All Bots for one user share the computer
- Cursor Cloud Agent
- No; each agent has a dedicated VM boundary
- Control question
- How are external actions checked?
- Grok Bot
- Task boundaries, approval cards, and model-based Auto Review
- Cursor Cloud Agent
- Repository scope, egress policy, hooks, and human review of the draft PR
- Control question
- How are credentials supplied?
- Grok Bot
- Human takeover, secure connection handoff, connectors, and MCPModel Context Protocol. A standard that lets an AI agent pull in context from outside the repo, like Jira tickets or internal docs. Press Enter for the full definition. policy
- Cursor Cloud Agent
- Encrypted secrets, runtime-secret redaction, build secrets, and OIDCOpenID Connect. The modern single sign-on standard, built as an identity layer on top of OAuth 2.0. Where SAML is XML and enterprise-legacy, OIDC is JSON and what newer tools implement first. Press Enter for the full definition. options
- Control question
- Is full run evidence available?
- Grok Bot
- Spend and usage now; Bot-action audit view is coming
- Cursor Cloud Agent
- Dashboard sessions, Git attribution, signed commits, audit events, and run diagnostics
- Control question
- Can admins set a product spend cap?
- Grok Bot
- No Grok BotAn early AI teammate product from SpaceXAI and Cursor that works across apps on a persistent cloud computer and asks for approval on selected actions; launch materials describe it as beta, but Cursor says Beta Services terms do not apply. Press Enter for the full definition.-specific cap yet; account controls apply
- Cursor Cloud Agent
- A spend limit is requested when Cloud AgentsAgents that run in a Cursor-managed virtual machine, check out the repo, do the work and open a pull request, then shut down, with no load on your laptop. Press Enter for the full definition. are first used
| Control question | Grok Bot | Cursor Cloud Agent |
|---|---|---|
| Can one run see another run's workspace? | All Bots for one user share the computer | No; each agent has a dedicated VM boundary |
| How are external actions checked? | Task boundaries, approval cards, and model-based Auto Review | Repository scope, egress policy, hooks, and human review of the draft PR |
| How are credentials supplied? | Human takeover, secure connection handoff, connectors, and MCPModel Context Protocol. A standard that lets an AI agent pull in context from outside the repo, like Jira tickets or internal docs. Press Enter for the full definition. policy | Encrypted secrets, runtime-secret redaction, build secrets, and OIDCOpenID Connect. The modern single sign-on standard, built as an identity layer on top of OAuth 2.0. Where SAML is XML and enterprise-legacy, OIDC is JSON and what newer tools implement first. Press Enter for the full definition. options |
| Is full run evidence available? | Spend and usage now; Bot-action audit view is coming | Dashboard sessions, Git attribution, signed commits, audit events, and run diagnostics |
| Can admins set a product spend cap? | No Grok BotAn early AI teammate product from SpaceXAI and Cursor that works across apps on a persistent cloud computer and asks for approval on selected actions; launch materials describe it as beta, but Cursor says Beta Services terms do not apply. Press Enter for the full definition.-specific cap yet; account controls apply | A spend limit is requested when Cloud AgentsAgents that run in a Cursor-managed virtual machine, check out the repo, do the work and open a pull request, then shut down, with no load on your laptop. Press Enter for the full definition. are first used |
A control existing on one product should not be assumed on the other.
Cloud AgentsAgents that run in a Cursor-managed virtual machine, check out the repo, do the work and open a pull request, then shut down, with no load on your laptop. Press Enter for the full definition. are narrower by design, which makes their authorization easier to state. An admin connects source control, grants selected repositories, and each user adds their own account. The agent cannot reach a repository that user could not reach. Grok BotAn early AI teammate product from SpaceXAI and Cursor that works across apps on a persistent cloud computer and asks for approval on selected actions; launch materials describe it as beta, but Cursor says Beta Services terms do not apply. Press Enter for the full definition. can cross websites and tools after the user signs in, so the source system's own scopes and revocation become more important.
Grok BotAn early AI teammate product from SpaceXAI and Cursor that works across apps on a persistent cloud computer and asks for approval on selected actions; launch materials describe it as beta, but Cursor says Beta Services terms do not apply. Press Enter for the full definition.'s current audit gap changes high-assurance use. The team docs say an action audit view is coming. If a process requires a complete, durable record of every action, keep the authoritative log in the source system or leave that process out of the early rollout. For code, Cloud Agent commits and pull requests already create an attributable trail, though a Git trail is not a record of every thought or tool action.
How do privacy, model choice, and billing compare?
Neither product supports Privacy Mode (Legacy)A Cursor data setting that blocks Grok Bot because the product requires cloud data storage. Press Enter for the full definition. because both need cloud storage. Standard Cursor privacy settings govern them, including the applicable training protections. Cloud Agent documentation publishes detailed retention by data type and model-by-model provider exceptions. Grok BotAn early AI teammate product from SpaceXAI and Cursor that works across apps on a persistent cloud computer and asks for approval on selected actions; launch materials describe it as beta, but Cursor says Beta Services terms do not apply. Press Enter for the full definition. documentation requires storage and points to Cursor's privacy terms for backend retention.
Model control diverges. Cloud AgentsAgents that run in a Cursor-managed virtual machine, check out the repo, do the work and open a pull request, then shut down, with no load on your laptop. Press Enter for the full definition. expose a curated model selection and supported context-window choices. Grok BotAn early AI teammate product from SpaceXAI and Cursor that works across apps on a persistent cloud computer and asks for approval on selected actions; launch materials describe it as beta, but Cursor says Beta Services terms do not apply. Press Enter for the full definition. offers no picker to a member or admin. The product routes a fixed set of models by surface with automatic failover, reports the model that served each request in usage analytics, and bills that actual serving model.
Cloud AgentsAgents that run in a Cursor-managed virtual machine, check out the repo, do the work and open a pull request, then shut down, with no load on your laptop. Press Enter for the full definition. charge at API pricing for the selected model, and Cursor asks for a spend limit at first use. Current Cursor Help includes Grok BotAn early AI teammate product from SpaceXAI and Cursor that works across apps on a persistent cloud computer and asks for approval on selected actions; launch materials describe it as beta, but Cursor says Beta Services terms do not apply. Press Enter for the full definition. with SuperGrok Heavy, Cursor Ultra, or an eligible Premium team seat; Pro, Pro Plus, and a Standard seat alone do not include it. Paid Grok Bot usage resets weekly. There is no Grok Bot-specific spend cap yet.
That makes a universal cost comparison misleading. A Cloud Agent's model is visible before the run, while a Grok BotAn early AI teammate product from SpaceXAI and Cursor that works across apps on a persistent cloud computer and asks for approval on selected actions; launch materials describe it as beta, but Cursor says Beta Services terms do not apply. Press Enter for the full definition. request may fail over to another managed model and bills accordingly. Compare the cost of one measured workflow under the actual account settings, and include the human review time each output needs.
When should you choose Grok Bot or Cursor Cloud Agents?
Choose from the final artifact and its owner. A task that ends at a branch or pull request belongs with a Cloud Agent. A role that wakes up tomorrow and needs the same apps, preferences, and recurring responsibility belongs with Grok BotAn early AI teammate product from SpaceXAI and Cursor that works across apps on a persistent cloud computer and asks for approval on selected actions; launch materials describe it as beta, but Cursor says Beta Services terms do not apply. Press Enter for the full definition.. Use both when the wider process has a clear repository subtask.
- Choose Grok BotAn early AI teammate product from SpaceXAI and Cursor that works across apps on a persistent cloud computer and asks for approval on selected actions; launch materials describe it as beta, but Cursor says Beta Services terms do not apply. Press Enter for the full definition. for research, reconciliations, source-linked briefs, document work, recurring preparation, and supervised actions across tools.
- Choose Cursor Cloud AgentsAgents that run in a Cursor-managed virtual machine, check out the repo, do the work and open a pull request, then shut down, with no load on your laptop. Press Enter for the full definition. for feature work, refactors, migrations, test repair, multi-repo engineering, and code changes that should return through Git.
- Choose both when a standing Bot detects or prepares a task, delegates the repository work, then incorporates the reviewed result into the wider workflow.
- Choose neither yet when the workflow requires an unavailable audit control, a blocked privacy mode, or source access you cannot scope and revoke.
A product-performance workflow shows the split. Grok BotAn early AI teammate product from SpaceXAI and Cursor that works across apps on a persistent cloud computer and asks for approval on selected actions; launch materials describe it as beta, but Cursor says Beta Services terms do not apply. Press Enter for the full definition. can review dashboards, traces, support threads, and release notes, then return the highest-confidence hotspot with evidence. If the next step is a patch, hand that bounded finding to a Cloud Agent with the repository, acceptance test, and no authority to merge. The Bot can then link the resulting pull request in its report.
Start with one handoff, not a chain. Test the Grok BotAn early AI teammate product from SpaceXAI and Cursor that works across apps on a persistent cloud computer and asks for approval on selected actions; launch materials describe it as beta, but Cursor says Beta Services terms do not apply. Press Enter for the full definition. result or the Cloud Agent pull request on its own, then connect them after each side has a review standard. A two-product workflow that fails in the middle is otherwise harder to diagnose than either product alone.
How can Grok Bot and Cursor Cloud Agents work together?
Grok BotAn early AI teammate product from SpaceXAI and Cursor that works across apps on a persistent cloud computer and asks for approval on selected actions; launch materials describe it as beta, but Cursor says Beta Services terms do not apply. Press Enter for the full definition. can launch Cursor cloud agents when the team-wide Cloud AgentsAgents that run in a Cursor-managed virtual machine, check out the repo, do the work and open a pull request, then shut down, with no load on your laptop. Press Enter for the full definition. toggle allows it. Keep the delegation narrow: give the Cloud Agent the repository task, expected tests, and review artifact, while the Bot retains the business context and waits for the result.
- 1Let Grok BotAn early AI teammate product from SpaceXAI and Cursor that works across apps on a persistent cloud computer and asks for approval on selected actions; launch materials describe it as beta, but Cursor says Beta Services terms do not apply. Press Enter for the full definition. gather current evidence and state the problem without changing production.
- 2Review the proposed coding task and its acceptance check.
- 3Have Grok BotAn early AI teammate product from SpaceXAI and Cursor that works across apps on a persistent cloud computer and asks for approval on selected actions; launch materials describe it as beta, but Cursor says Beta Services terms do not apply. Press Enter for the full definition. launch a Cursor Cloud Agent for that repository task.
- 4Review the Cloud Agent's branch, tests, artifacts, and draft pull request through the normal code process.
- 5Return the approved result to the Grok BotAn early AI teammate product from SpaceXAI and Cursor that works across apps on a persistent cloud computer and asks for approval on selected actions; launch materials describe it as beta, but Cursor says Beta Services terms do not apply. Press Enter for the full definition. thread for the wider report or next supervised action.
Do not let the handoff collapse the two approval systems. A Grok BotAn early AI teammate product from SpaceXAI and Cursor that works across apps on a persistent cloud computer and asks for approval on selected actions; launch materials describe it as beta, but Cursor says Beta Services terms do not apply. Press Enter for the full definition. approval to launch work is not approval to merge the returned code, and a green pull request does not authorize the Bot to send a customer message about it. Each product should stop at its own consequential boundary.
For setup and a safe first handoff, use the Grok BotAn early AI teammate product from SpaceXAI and Cursor that works across apps on a persistent cloud computer and asks for approval on selected actions; launch materials describe it as beta, but Cursor says Beta Services terms do not apply. Press Enter for the full definition. tutorial. For the user-scoped computer, privacy requirement, and current admin gaps, read the security guide before connecting production systems.
Frequently asked questions
Is Grok Bot a replacement for Cursor Cloud Agents?
No. Grok Bot owns persistent cross-app work, while Cursor Cloud Agents own bounded repository tasks in isolated development VMs. Grok Bot can launch a Cloud Agent for a coding step when the team permits it.
Which is better for writing and testing code?
Cursor Cloud Agents are the better starting surface. They clone authorized repositories, run in configured development environments, test changes, push a branch, and can open a draft pull request with artifacts for review.
Which is better for recurring work across business apps?
Grok Bot. Its named role, conversation, files, browser sessions, preferences, skills, and routines persist across work, making it suited to recurring research, reconciliation, preparation, and supervised actions.
Do Grok Bot and Cloud Agents use the same VM model?
No. All Bots for one user share one managed computer. Every Cursor Cloud Agent run gets its own isolated VM boundary. The different workspace design is central to their security and persistence models.
Can I choose the model in both products?
No. Cloud Agents offer a curated model selection and supported context sizes. Grok Bot manages its own fixed model set and automatic failover, with no user or admin model picker.
Can Grok Bot launch a Cursor Cloud Agent?
Yes. The Grok Bot team setup includes a Cloud Agents toggle. When enabled, a Bot can delegate a repository coding task to a Cloud Agent and continue coordinating the wider workflow.
Sources & last verified
- SpaceXAI Docs: Grok Bot overview
- SpaceXAI Docs: Grok Bot for teams and enterprises
- SpaceXAI Docs: Approvals, security, and privacy
- Cursor Docs: Cloud Agents
- Cursor Docs: Cloud Agent security overview
- Cursor Docs: Cloud environment setup
Cursor ships frequently. Facts verified against primary sources on August 14, 2026.