Cursor logo

Cursor

Editor reviewed

AI coding agent and VS Code-based editor for multi-file edits and cloud agents

5.0(1 review)Freemiumcursor.comUpdated
Visit website

About Cursor

Cursor is an AI coding tool from Anysphere that combines a desktop code editor with autonomous coding agents. The editor is built on VS Code, so existing extensions, themes and keybindings carry over. Developers use it for inline edits (Cmd+K), Tab completions, chat that references files and the indexed codebase, and agents that carry out multi-file tasks and show the changes as diffs for review.

Best for

  • Multi-file feature work, codebase-wide refactoring, delegating tasks to background agents, reviewing pull requests
  • Feature implementation: Ask an agent to add a route, migration and UI component across several files and review the diff.
  • Refactoring and migrations: Use codebase search and agents to update deprecated patterns across a repository.
  • Background tasks: Start cloud agents from Slack or the web to fix issues or open pull requests while you work on something else.
  • Pull request review: Run Bugbot on GitHub pull requests to catch bugs before merge.
  • Codebase onboarding: Ask questions about an unfamiliar repository using indexed search.

Not ideal for

  • Non-coding tasks
  • Users who want predictable flat-rate pricing with third-party models included

Popular with

  • Software engineers
  • Full-stack developers
  • Engineering teams
  • Enterprises

Screenshots

Key features

  • Agents
  • Cloud agents
  • Tab and inline edits
  • Codebase indexing
  • Model choice
  • Bugbot code review
  • CLI
  • Slack app and iOS app
  • Marketplace and rules
  • Privacy mode

Pros and cons

Pros

  • Familiar VS Code-based editor with extension and keybinding support
  • Local and cloud agents that can run in parallel on multi-file tasks
  • Choice of models from several providers, including Cursor's own
  • Covers editor, CLI, web, Slack, iOS and PR review in one product
  • Privacy mode and SOC 2 and ISO certifications for team use

Cons

  • Third-party model usage is billed separately from the plan at API list price
  • Pricing is bundled with Grok and uses weekly usage pools, which makes costs harder to predict
  • No Android app, and the iOS app only controls local agents
  • Indexing can be heavy on very large repositories

Pricing

Plans as listed by the vendor on . Check their site for current prices.

See cursor.com

Free

$0

Limited usage for trying Cursor and Grok.

  • Limited usage
  • Limited chat requests
  • Limited agent requests

Plus

$8/month

Basic usage pool, Grok models only.

  • Basic shared usage pool
  • Grok models only
  • Cursor agent
  • Grok API access

Premium

Most popular

$30/month

4x the usage of Plus, recommended for new AI users.

  • 4x Plus usage
  • Latest Grok models
  • Cloud agents and automations
  • Bugbot code reviews
  • Third-party models at API list price

Super

$100/month

15x the usage of Plus, for daily agent users.

  • 15x Plus usage
  • Everything in Premium
  • Priority model access at peak times

Ultra

$200/month

50x the usage of Plus, for power users.

  • 50x Plus usage
  • Everything in Super
  • Earliest access to new products and models

Teams

$30/user/mo

Per-seat plans with pooled admin controls; seats are Premium $30, Super $100 or Ultra $200 per user per month.

  • Premium $30
  • Super $100 or Ultra $200 seats
  • Weekly usage pool per seat
  • Centralized billing
  • Usage analytics
  • Team privacy settings
  • SAML/OIDC SSO
  • Team marketplace

Enterprise

Custom/custom

Invoicing, pooled usage and advanced security.

  • Invoice and PO billing
  • SCIM
  • Audit logs
  • Granular spend controls
  • Repository and MCP access controls
  • Priority support

Our ratings

Scored out of 10 by the Finder Atlas editorial team.

8.7overall
Inline Code Completion
9.2
Codebase Indexing & Context
9.0
Multi-file Agentic Edits
9.3
IDE & Extension Compatibility
9.0
Pricing Clarity
6.8

Platforms and integrations

Available on

  • macOS
  • Windows
  • Linux
  • Web
  • iOS
  • CLI

Works with

  • GitHub
  • Slack
  • VS Code extensions
  • MCP servers

Cursor alternatives

Other tools people compare with Cursor.

All 8 alternatives
  • OpenAI Codex logo
    OpenAI CodexFreemium

    OpenAI's coding agent that writes, reviews and runs code from terminal, IDE, desktop or cloud

    Compare
  • Conductor logo
    ConductorFreemium

    Run a team of coding agents in parallel, each in its own isolated workspace

    Compare
  • Macroscope logo
    MacroscopeFreemium

    AI code review, agent orchestration and engineering status updates for GitHub teams

    Compare
  • Cline logo
    ClineFreemium

    Open-source autonomous coding agent for your editor, terminal, and CI pipelines

    Compare
  • Augment Code logo

    Agentic coding platform that runs code review and ticket-to-PR work on your codebase

    Compare

Release history

  1. Cloud agents can build interactive canvases, a beta Environments API was added, and environments support up to 300 re...

    Cloud agents can build interactive canvases, a beta Environments API was added, and environments support up to 300 repositories.

  2. The iOS app can view and message local agents running on a computer

    The iOS app can view and message local agents running on a computer.

  3. New bots monitor deployments for regressions and report exploitable bugs on pull requests

    New bots monitor deployments for regressions and report exploitable bugs on pull requests.

  4. Beta feature for managing larger bodies of work with a coordinator agent that delegates tasks

    Beta feature for managing larger bodies of work with a coordinator agent that delegates tasks.

Reviews

5.0

Based on 1 review

  • FinderAtlas Curator
    2026-08-05
    5 out of 5

    Cursor has completely doubled my coding speed. The @codebase search combined with Claude 3.5 Sonnet is unreal.

Frequently asked questions

Is Cursor based on VS Code?

Yes. The editor is built on VS Code, so extensions, themes and keybindings from VS Code can be used.

Is my code used to train models?

Cursor says that with privacy mode enabled, which can be set in settings or by a team admin, code data is not used for training by Cursor or its model providers.

Are third-party models included in the subscription?

No. Plans include Grok model usage, while third-party models such as those from OpenAI or Anthropic are billed at API list price. You can set spending caps.

Does Cursor have a public API?

Yes. A Cloud Agents API (beta), TypeScript and Python SDKs, and Enterprise-only Admin and Analytics APIs. These run agent workflows and are not a standalone chat-completions API.

Is there a mobile app?

There is an iOS app that can view and message local agents running on your computer, which must stay on and online. No Android app is listed.

Who owns Cursor?

Cursor is made by Anysphere, which SpaceX agreed to acquire in June 2026. Cursor's site notes it is now part of SpaceX.

More Coding AI tools

View all