Cursor

A code editor where you describe the change you want in plain English and it edits the files directly, used by non-developers for small internal tools.

Websites & apps Harness

Checked 25 Sep 2026

Open cursor.com

What you'd use it for

Building or changing a small internal tool, script, or website by describing the edit in plain English inside a full code editor, without hiring a developer.

Ask the Agent to explain an unfamiliar codebase and point you to the files worth reading before you touch anything
Describe a feature or UI change in plain English and get a multi-file edit applied directly, with a diff to review before it lands
Keep a cloud agent working on a task in the background after you close the laptop
Run automatic code review on a pull request instead of reading every diff by hand
Switch between frontier models from different providers inside the same editor without changing tools

The verdict

Cursor is the strongest choice for someone who wants an entire editor built around an AI agent rather than a plugin bolted onto one they already use, and the quickstart genuinely gets you from install to a working change in minutes. The catch is the pricing model -- usage draws from a shared credit pool that different models burn through at different rates, so the headline monthly figure understates what a heavy user actually spends, and automatic code review on a pull request draws from that same pool rather than being a flat inclusion. It suits someone willing to sit inside a real code editor -- accepting diffs, managing context, occasionally reading a stack trace -- not someone who wants a finished, hosted app to appear from a one-line description with nothing else to manage.

What it costs

Tier	Price	For	Includes
Hobby	$0	The curious, trying it before committing.	Limited Agent requests, access to Composer, no credit card required.
Start (India only)	₹649/month	Developers in India who want a cheaper daily-driver plan; not sold outside India.	Non-fast access to Grok 4.7, Grok 4.6, Grok 4.5 and Composer 2.5 at a fixed medium effort, plus Cloud Agents, Cursor for iOS, and plugins, MCP servers, hooks and skills. Excludes third-party models, on-demand usage, Bugbot, Auto, Automations and the Cursor SDK.
Pro	$20/month	One person coding daily and wanting frontier models on tap.	Everything in Hobby, plus extended Agent limits, access to frontier models, Grok Bot access, MCPs, skills and hooks, Cloud Agents, and Bugbot code review billed from usage.
Pro Plus	$60/month	A daily agent user who needs a higher usage ceiling than Pro.	Everything in Pro, plus 3x Pro's Agent limits and higher Grok Bot usage limits.
Ultra	$200/month	An agent power user running multiple agents or heavy automation.	Everything in Pro, plus 20x Pro's Agent limits, the highest Grok Bot usage limits, and priority access to new features.
Teams Standard	$40/user/month	Teams that ship together and need shared billing and context.	Everything in Pro, plus centralised team billing and administration, an internal marketplace for rules, skills and plugins, cloud agents and automations with shared team context, agentic code review with Bugbot billed from on-demand spend, usage analytics, team-wide privacy mode, and SAML/OIDC SSO.
Teams Premium	$120/user/month	Teams whose agent usage regularly outgrows the Standard seat.	Everything in Teams Standard, plus 5x the Standard seat's Agent limits.
Enterprise	Custom pricing	Scaled organisations with governance requirements.	Everything in Teams, plus pooled usage, invoice/PO billing, SCIM seat management, repository, model and MCP access controls, auto-run, browser and network controls, audit logs, service accounts, and an AI code-tracking API.

Read from Cursor's own pricing page on 25 Sep 2026. Vendors change prices without notice.

What you need to start

A free account, created by signing in inside the app -- no credit card needed for the free Hobby tier
macOS 12 or later, Windows 10 or later, or a supported Linux distribution (Debian, Ubuntu, RHEL, Fedora, or AppImage)

A few minutes from download to the Agent's first codebase summary, per Cursor's own quickstart; getting comfortable reviewing diffs and managing context for real changes takes longer than the install does.

Your first run Download the desktop app, sign in, and open a project folder; the first real result is asking the Agent (Cmd/Ctrl I) to explain the codebase, which it answers by reading the relevant files and summarising how the project fits together in under a minute.

Where it breaks

Bugbot, Cursor's automated code-review feature, is billed on usage on top of the seat price rather than included in it, which is easy to miss when comparing the headline numbers to a flat-rate competitor.

A desktop app for macOS, Windows and Linux, built on a fork of VS Code — existing extensions and keybindings mostly carry over.

First entry in this reference. There is no previous check to compare it against.

When not to use it

Not for someone who wants a hosted app to just appear from a prompt with hosting and a URL handled for them -- that is a builder like Lovable or Bolt.new's job, not an editor's. Also not the right fit for someone unwilling to open a real code editor or review a diff before it is applied, since the whole product assumes you are sitting inside one, even if you never write a line of code yourself.

Cursor is what happens when you stop bolting AI onto an existing editor and build the editor around AI instead. It looks like VS Code because it is one, underneath — but the chat panel, the inline edits, and the agent that can work across your whole project are the default way of writing code here, not an add-on.

The $20 Pro tier is squarely aimed at one person who wants Cursor as their daily driver: extended agent limits, access to whichever frontier model is current, Bugbot code review on a pull request billed from usage, and the cloud-agent feature that lets a task keep running after you close the laptop. Pro Plus ($60) and Ultra ($200) raise the Agent usage ceiling for heavier daily use. Teams adds the coordination layer on top of that same feature set — centralised billing and administration, shared team context for cloud agents, usage analytics, and SSO — not automatic code review, which Pro already includes.

Who reaches for this over the alternatives

Developers are the obvious audience, but Cursor shows up in non-developer hands surprisingly often — someone building a single internal script, a scraper, a small dashboard, who would never open a terminal willingly but will describe what they want in English and watch it appear. That is the same use case GitHub Copilot and Claude Code compete for, and the honest answer is that the three overlap heavily; the difference is mostly which one is already wired into your existing workflow — GitHub-native teams lean Copilot, terminal-first developers lean Claude Code, and Cursor wins the people who want the whole editor rebuilt around the assistant.

Where it stops

Cursor acts on your files when you ask it to, inside an editor you have open. It does not run unattended on a schedule — that job belongs to something like Make triggering a headless build.