Cursor
Cursor is an AI-first code editor built on a VS Code fork, featuring agent mode for autonomous multi-file edits, predictive tab completion, and cloud agents for running long tasks in isolated environments.
What is Cursor?
Cursor is an AI-first code editor developed by Anysphere, built on a fork of VS Code. Its official positioning is as a "coding agent for building ambitious software," and the product is designed around agent-driven development rather than AI as a supplementary autocomplete layer[reference:0]. The editor supports macOS, Windows, and Linux, and offers one-click import of VS Code settings, keybindings, themes, and extensions, keeping the default VS Code shortcuts for a familiar transition[reference:1].
Cursor is now sold as part of The X Subscription, a combined offering that includes Cursor, Grok, and X. Existing subscribers on legacy Cursor plans (Hobby, Pro, Pro+, Ultra, Teams Standard/Premium) can continue using their plans, but new customers select a plan on The X Subscription[reference:2].
How does Cursor work?
Users interact with Cursor through natural-language instructions in the editor. The primary workflow centers on Agent mode: a user describes a task, and the agent plans and implements changes across multiple files, exploring the codebase, searching for relevant context, and making edits autonomously. Run modes control how much the agent may do without approval, and developers review the diff or pull request before merging[reference:3].
For tasks that benefit from isolation or extended runtime, Cloud Agents run in dedicated cloud VMs, clone the repository, execute install scripts, and push signed commits back. Local agent sessions can be handed off to Cloud Agents to continue running while the developer is away, with results accessible via the web or mobile[reference:4].
Cursor also provides predictive Tab completions, inline Cmd+K edits, and an Agents Window for managing running agent sessions[reference:5]. The output is surfaced directly in the editor as text, code changes, diffs, or pull requests, which the developer can review, refine, or reject.
Use Cases
- Building features across multiple files and modules with agent-driven implementation
- Delegating long-running refactoring or dependency-update tasks to Cloud Agents that continue in the background
- Understanding and navigating unfamiliar codebases using codebase-aware chat and agent exploration
- Accelerating repetitive editing patterns with predictive Tab completions
- Prototyping and iterating on software with a mix of inline edits and autonomous agent execution
- Reviewing and improving code quality with Bugbot code review on Premium-tier plans and above
Pricing
Pricing model: Freemium
| Plan | Price | Billing | Included usage | Notes |
|---|---|---|---|---|
| Free | $0 | per month | Limited Agent and Chat usage with Cursor models. Third-party models are not available on the Free plan. | — |
| Plus (The X Subscription) | $8 | per month | Base usage of Cursor models (Grok and Composer) at fixed settings. Does not include Cloud Agents, Automations, third-party models, or your own API keys. | — |
| Premium (The X Subscription) | $30 | per month | 4x the usage of Plus. Adds Cloud Agents, Automations, Bugbot code review, Origin code hosting, and third-party models billed at API list price from prepaid credit. | — |
| Super (The X Subscription) | $100 | per month | 15x Plus usage with priority model access at peak times. | — |
| Ultra (The X Subscription) | $200 | per month | 50x Plus usage and earliest access to new products and models. | — |
Common Limitations
- Plan and model access · The Free plan offers limited Agent and Chat usage with Cursor models only. The $8 Plus plan on The X Subscription also excludes Cloud Agents, Automations, third-party models (e.g., Claude, GPT), and your own API keys. Access to third-party models and advanced agent features requires the Premium tier ($30/mo) or above.
- Third-party model billing · On Premium and higher plans, third-party models are not included in the subscription price. They are billed at API list price from prepaid credit. Legacy plans (Pro, Pro+, Ultra, Teams) use separate usage pools for Cursor and other models.
- Privacy Mode default · Privacy Mode is available on all plans, but it is off by default for individual accounts. When disabled, Cursor states it may store and use codebase data, prompts, editor actions, and code snippets to improve the product. Teams and Enterprise admins can enforce it organization-wide.
- Cloud Agents and infrastructure · Cloud Agents, which run in isolated cloud VMs, are available on Premium and above on The X Subscription, and on legacy Pro, Pro+, Ultra, and Teams plans. They are not available on the Free or Plus plans. Data residency options such as US-only storage are Enterprise-only.
- Editor dependency · Cursor is a standalone desktop application built on a VS Code fork for macOS, Windows, and Linux. It is not a plugin for an existing IDE. Users must migrate their development environment, though Cursor offers one-click import of VS Code settings, keybindings, themes, and extensions.
Key Features
- Agent mode for autonomous multi-file edits
- Predictive Tab completions
- Inline Cmd+K editing
- Cloud Agents running in isolated VMs
- Codebase-aware chat and context exploration
Cursor vs. GitHub Copilot, Claude Code, and Devin Desktop (Windsurf)
This comparison uses official product and pricing information available as of November 2026. It focuses on documented positioning, plan structures, and feature access boundaries rather than subjective assessments of code-generation quality. Pricing and plan features are subject to change.
| Tool | Positioning | Key differences | Best fit | Pricing note |
|---|---|---|---|---|
| Cursor Official product page Pricing and plans (docs) Models & Pricing (docs) | An AI-first code editor built on a VS Code fork, positioning an agent at the center of the development workflow. | Offers Agent mode for autonomous multi-file edits, Cloud Agents in isolated VMs, and predictive Tab completions. Now sold as part of The X Subscription, with a $0 Free tier, an $8 Plus tier that excludes third-party models and Cloud Agents, and Premium ($30) and above tiers that add those capabilities. | Developers who want an editor-native agent experience with autonomous multi-file editing and the option to run long tasks in the cloud. | Free ($0, limited). The X Subscription: Plus $8/mo, Premium $30/mo, Super $100/mo, Ultra $200/mo. Legacy Cursor plans (Pro $20/mo, etc.) remain for existing subscribers but are not offered to new customers. |
| GitHub Copilot Official plans page Copilot feature overview | An AI coding agent that helps developers write, understand, and ship software across the GitHub workflow. | Integrated across VS Code, Visual Studio, JetBrains, Xcode, Vim/Neovim, Eclipse, the terminal/CLI, GitHub.com, mobile, and Slack/Teams. Built into GitHub's issue, pull request, and code review workflow. Offers a Free tier with 2,000 completions/month, Pro ($10/mo), Pro+ ($39/mo), and Max ($100/mo). | Developers and teams already working within the GitHub ecosystem who want an AI agent integrated across multiple surfaces without switching editors. | Free plan available with limited chat and agent usage. Individual paid plans: Pro $10/user/mo, Pro+ $39/user/mo, Max $100/user/mo. Business and Enterprise plans are also available. |
| Claude Code Claude pricing page Anthropic announcement on Claude Code and business plans | An agentic coding tool from Anthropic that reads the codebase, edits files, runs commands, and integrates with development tools. | Available across the terminal, VS Code, JetBrains, Slack, the web, desktop, iOS, and Android from a single subscription. Works as both an MCP client and an MCP server. There is no standalone Claude Code plan or free tier; it shares a usage pool with the Claude subscription and is available on Claude Pro, Max, Team, and Enterprise plans. | Developers who prefer an agentic coding tool that spans terminal, IDE, and mobile surfaces under the Claude subscription, particularly those already using Anthropic's models. | No free tier for Claude Code. Available on Claude Pro ($20/mo monthly or $17/mo annual), Max, Team, and Enterprise plans. Usage draws from the same pool as chats on the Claude subscription. |
| Devin Desktop (formerly Windsurf) Devin Desktop FAQ (official) Devin pricing (official) | An AI-powered IDE originally built by Codeium, now positioned as the IDE surface for Cognition's Devin platform. | Formerly known as Windsurf, the product was renamed Devin Desktop on June 2, 2026, after Cognition's acquisition. It features Cascade, an agentic AI for multi-file edits and terminal commands, and retains a VS Code fork foundation. Pricing is shared with the Devin platform: a free plan, Pro at $20/mo, Max at $200/mo, and Teams starting at $80/mo. | Developers looking for an agentic IDE experience with a free tier and a $20/mo Pro plan, particularly those interested in the Devin platform ecosystem. | Free plan available. Pro $20/mo, Max $200/mo, Teams from $80/mo (with full seats at $40 each). Enterprise pricing is custom. |
Comparison checked against official product information on 2026-11-03. Features and pricing may change.
Frequently Asked Questions
Is Cursor free to use?
Cursor has a Free plan ($0) with limited Agent and Chat usage on Cursor models. Third-party models and advanced features such as Cloud Agents and Automations are not included. Paid plans on The X Subscription start at $8/month for Plus, but the Plus plan also excludes third-party models and Cloud Agents.
What is the difference between Cursor's Free, Plus, and Premium plans?
The Free plan offers limited usage of Cursor models. The $8 Plus plan provides base usage of Cursor models at fixed settings but excludes Cloud Agents, Automations, third-party models, and your own API keys. The $30 Premium plan adds 4x Plus usage, Cloud Agents, Automations, Bugbot code review, Origin code hosting, and access to third-party models billed at API list price from prepaid credit.
Are third-party models like Claude and GPT included in Cursor's subscription?
No. On The X Subscription, third-party models are not included in the plan price. They are available on Premium and above, billed at API list price from prepaid credit. On legacy Cursor plans, third-party models use a separate usage pool.
Does Cursor replace VS Code or work alongside it?
Cursor is a standalone desktop application built on a VS Code fork, not a VS Code extension. It offers one-click import of VS Code settings, extensions, themes, and keybindings, and retains VS Code's default shortcuts, but users must switch to the Cursor application to use its AI features.
What are Cursor Cloud Agents and which plans include them?
Cloud Agents run Cursor's agent in isolated cloud VMs, cloning the repository and executing tasks such as installs, builds, and code changes without tying up local resources. They are available on Premium and above on The X Subscription, and on legacy Pro, Pro+, Ultra, and Teams plans. They are not included on the Free or $8 Plus plans.
Alternative Tools
-
GitHub Copilot GitHub Copilot is an AI coding agent embedded across GitHub, IDEs, the terminal, and the web. It combines unlimited completions on paid plans with agent mode, asynchronous coding agents, and code review, but its value depends on disciplined review, awareness of IP and data-training policies, and realistic expectations about PR acceptance rates and context limits. -
Ollama Ollama is an open-source local LLM runtime and model manager that makes running open-weight models on your own hardware a single command. Local inference is free and unlimited; the costs are hardware, VRAM-gated context defaults, and a documented security surface that requires deliberate network isolation. -
OpenAI Codex OpenAI Codex is an agentic coding product bundled into ChatGPT plans, spanning CLI, IDE, desktop, and cloud. Plus at $20/month is the real entry point; Free and Go are demos. Usage is credit-metered with wide ranges, ChatGPT and Codex pools merge in December 2026, and disclosed security boundary issues remain relevant for teams evaluating it for sensitive codebases. -
Claude Code Claude Code is Anthropic's agentic coding tool that runs in the terminal, IDE, and desktop, powered by Sonnet and Opus models, with checkpointing, subagents, hooks, MCP support, and a native VS Code extension—available only through paid Claude subscriptions or API access, with no free tier. -
Bolt.new Bolt.new is StackBlitz's in-browser AI full-stack app builder. It runs a Node.js runtime inside your browser tab, installs packages, starts dev servers, and generates full-stack code from prompts—no local environment setup required. -
Lovable Lovable is an AI full-stack app builder that turns natural-language descriptions into working web applications with frontend, backend, database, auth, and deployment. It is unusually effective at getting to a functional prototype and unusually poor at producing secure production code without human intervention; third-party scans find security issues in 99% of Lovable projects, and the managed backend creates migration friction that the code-portability narrative does not fully offset.