The short answer
Cursor is a code editor with AI built in. It is based on VS Code, so anyone who knows that editor will recognise the layout, the shortcuts and the extensions. The AI is what sets it apart: Cursor completes your code as you type, answers questions about your project and can run a whole task with an agent, from editing files to running commands.
Cursor is made by Anysphere, a San Francisco company founded in 2022 by four MIT students. SpaceX acquired it in 2026. It is one of the most widely used AI tools among developers.
What Cursor does
You work in Cursor the way you work in any editor: open your project and write code. The AI joins in at three levels. It finishes your line, it answers your question, or it takes over the whole task.
Completing as you type
Cursor predicts what you are about to write, often several lines at once, and even the next place you will need to change. Press Tab to accept the suggestion.
Questions about your code
Ask a question in a chat panel next to your code, such as where invoices are calculated or why a test fails. Cursor looks for the answer in your own project.
An agent for whole tasks
In agent mode you give an instruction in plain language. Cursor finds the files involved, edits them, runs commands in the terminal and shows you what changes in each file.
Agents in the cloud
A task can also run on Cursor's servers while you carry on working. That way several tasks run at once, and you review the results later.
Conventions as rules
Rules are files in your project where you write down the style you follow, the commands that exist and what the AI must never do. Cursor passes them to the model with every request.
Cursor works with models from several providers. You pick one per task from Claude, GPT, Gemini and Grok among others, and Cursor also has a model of its own, Composer. That makes Cursor an AI harness: the software around a language model that gives it tools, context and limits. What that means is in what is an AI harness.
What does Cursor cost?
There is a free version with a limited number of AI requests, enough to get to know the editor. Paid plans for individuals run from 8 to 200 dollars a month, and teams pay per user. The more expensive the plan, the more AI usage it includes.
Two things to watch. The cheapest plan only gives access to the Grok models. And usage of models from other providers, such as Claude and GPT, is billed separately on top of the monthly price, at the provider’s own rates. If you mainly want to work with Claude or GPT, expect to pay more than the plan itself. Prices change often, and the current ones are on cursor.com.
Cursor, Claude Code and Lovable compared
All three let a language model write code. The difference is who they are built for and how much you still do yourself.
| Cursor | Claude Code | Lovable | |
|---|---|---|---|
| What it is | A code editor with AI built in, based on VS Code. | A coding agent that carries out tasks on its own in your project. | A tool that builds a complete web app from a description in a chat. |
| Built for | Developers who write code themselves. | Developers who hand off tasks and review the result. | People who don't write code: founders, product managers, business owners. |
| Where you work | In the editor itself, on your own machine. | Terminal, VS Code, JetBrains, desktop app or browser. | In the browser, with a chat on the left and your app on the right. |
| Who writes the code | You and the AI together, or the agent when you hand off a task. | The agent. You give the brief and judge the result. | Lovable. You only see the code if you go looking for it. |
| Models | A choice of models from several providers, plus its own. | Anthropic's Claude models. | Lovable picks the model for you. |
| Fits | Day-to-day programming in an existing project. | Tasks that touch many files, including in large codebases. | Testing an idea fast with a working prototype. |
The lines do blur. Cursor has an agent that runs whole tasks, and Claude Code runs as an extension inside your editor, Cursor included. That is why plenty of developers use both. Lovable sits further away: if you don’t read code, an editor has little to offer you. More on those two is in what is Claude Code and what is Lovable, and a wider look at the tools in vibe coding tools.
How a senior team works with Cursor
Our developers use Cursor alongside Claude Code. What makes the difference is how you put the AI to work and what you do with what it produces.
Rules for every project
A good set of rules covers the conventions, the folder structure, the test commands and what is off limits. Every mistake the AI makes twice becomes a rule.
The right model for the task
A fast, cheap model for completions and small questions, a stronger one for a tricky refactor or a plan. That saves money and produces better code.
Small briefs for the agent
One function, one bug, one screen at a time. A small change takes a few minutes to review. A change across forty files does not.
Tests as a safety net
Have the agent run the tests after every change. Without tests you are relying on what the AI tells you, and that is not enough.
Read every change
What the AI writes goes through the same review as a colleague's work. On security, payments and customer data a senior developer looks extra hard.
Cursor makes typing almost free. That turns reviewing code into the work that counts.
That takes experience. A suggestion from the AI nearly always looks right, even when it misses an edge case or opens a security hole. How to catch that across a whole project, from first setup to production, is in how to develop software with AI.
Where Cursor is a weaker fit
Cursor is an editor for people who read code. Without that knowledge you can’t judge whether a change is right, even when you can see exactly what changed. If you want to build an app without programming knowledge, a tool like Lovable is an easier start, and have the result checked before customers use it.
Large projects with many moving parts also need guidance. The agent doesn’t always see how everything connects, and a change in one place can break something in another. Tests and good rules catch that. And if you would rather hand a task off completely, from the terminal or in the background, an agent like Claude Code is often the better way of working. Which tool fits which situation is in the best AI harness for coding.
Conclusion: an editor that writes along with you
Cursor brings AI to the place where developers already work. You type less, find your way around a project faster and hand whole tasks to an agent. The work that remains is the work that takes experience: giving the right brief and judging what comes back.
Cursor = writing faster
Completions, questions and handing off tasks, in an editor you already know.
Senior developer = judgement
The rules, the choice of model and the review. That is where quality comes from, and it stays with a person.
What is Cursor?
Is Cursor free?
What is the difference between Cursor and VS Code?
What is the difference between Cursor and Claude Code?
What is the difference between Cursor and Lovable?
Is Cursor safe for company code?
Build faster with AI without lowering the bar?
Our senior developers work with Cursor and Claude Code every day. Bring them onto your project, or have your team learn to use them well in a workshop.