Part of the Cursor AI Complete Guide (2026) series. Learn when to run Agents locally vs in the cloud — and how to hand off work without blocking your machine.

What Are Cloud Agents?

Cloud Agents extend Cursor beyond your local machine. Instead of waiting for a long-running task to complete on your computer, you hand work off to a cloud-based agent that continues independently.

Examples:

  • Large codebase refactoring
  • Documentation generation across many files
  • Pull request reviews
  • Long-running debugging sessions
  • Repository-wide analysis

This allows developers to continue working elsewhere while the Agent completes background tasks.

Local Agent vs Cloud Agent

Local AgentCloud Agent
Runs on your computerRuns remotely
Immediate interactionBackground execution
Faster feedback loopBetter for long-running work
Local files availableRemote execution environment
Best for interactive editingBest for automation & batch work
graph TB
    subgraph Local
        L[Local Agent] --> LE[Instant Edits]
        L --> LT[Terminal on Your Machine]
    end
    subgraph Cloud
        C[Cloud Agent] --> CB[Background Progress]
        C --> CR[Remote Repo Clone]
    end
    D[Developer] --> L
    D -->|Handoff| C

For interactive coding, the local Agent is usually the better choice. For well-defined tasks that may take longer, Cloud Agents improve productivity by continuing work asynchronously.

When to Use Cloud Agents

Good fits:

  • “Generate JSDoc for all exported functions in src/"
  • "Review this PR and post a structured summary"
  • "Refactor module X to use the new API client"
  • "Analyze test coverage gaps and propose new tests”

Poor fits:

  • Tight edit-debug loops on UI
  • Tasks requiring frequent human decisions mid-stream
  • Work on uncommitted local-only experiments not pushed to git
flowchart TD
    Q{Task well-defined?}
    Q -->|Yes| L{Long-running?}
    Q -->|No| LOCAL[Use Local Agent]
    L -->|Yes| CLOUD[Use Cloud Agent]
    L -->|No| LOCAL

Handoff Workflow

Recent Cursor updates support seamless handoff from the CLI using the & prefix — start locally and continue in the cloud when appropriate.

sequenceDiagram
    participant Dev as Developer
    participant Local as Local Agent
    participant Cloud as Cloud Agent
    Dev->>Local: Start task interactively
    Local->>Dev: Partial progress / plan
    Dev->>Cloud: Hand off with &
    Cloud->>Cloud: Continue in background
    Cloud->>Dev: Notification when complete
    Dev->>Dev: Review results

Always review cloud-generated changes before merging — same as local Agent output.

Cloud Agents and Git

Cloud Agents typically work against repository state available in the remote environment. Best practices:

  • Push feature branches before starting large cloud jobs
  • Use clear branch names so results are easy to find
  • Run CI on cloud-generated PRs
  • Never merge without local review and tests

Security & Privacy

Cloud Agents process code in Cursor’s remote infrastructure. For sensitive codebases:

  • Enable Privacy Mode
  • Confirm your plan’s data retention policies
  • Use enterprise governance features for team-wide enforcement
  • Avoid sending secrets in prompts — use environment variables in the remote environment

See the Cursor AI Complete Guide for the full security overview.

Combining Cloud Agents with Rules and MCP

Cloud Agents respect the same Project Rules and AGENTS.md as local Agents. If MCP is configured for the cloud environment, cloud jobs can also use GitHub, issue trackers, and deployment tools.

Recommended stack:

  1. Define Rules before starting cloud jobs
  2. Write a specific, scoped prompt
  3. Hand off to Cloud Agent
  4. Review diff + run tests locally
  5. Merge via normal PR process

Best Practices

  • Scope tasks narrowly — one feature or module per cloud job
  • Include acceptance criteria in the prompt (“all tests must pass”)
  • Use Plan mode locally first for ambiguous work, then cloud for execution
  • Monitor notifications so finished jobs do not sit unreviewed
  • Do not treat cloud output as production-ready without human review

Frequently Asked Questions

What is a Cursor Cloud Agent?

A background AI agent that runs on Cursor’s infrastructure to complete longer tasks while you work on other things.

Do Cloud Agents cost extra?

Usage counts against your plan’s Agent allowance. Check current pricing for Cloud Agent specifics.

Can I use Cloud Agents from the CLI?

Yes — including handoff patterns with the & prefix where supported in your Cursor version.

Are Cloud Agents the same as Bugbot?

No. Bugbot focuses on automated PR review; Cloud Agents are general-purpose background Agents for varied tasks.

Last reviewed: July 2026