010 Coding Collective

010 Coding Collective

What is Cursor? The AI code editor explained, what it costs and when it fits

Cursor is a code editor with AI built in, based on VS Code. What it does, what it costs, how it differs from Claude Code and Lovable, and how a senior team works with it.

AI Cursor AI-Assisted Coding AI Agents VS Code Claude Code Lovable Software Development

AI inside your own editor

Cursor completes your code as you type, answers questions about your project and can take on whole tasks with an agent. Work gets faster, and so do mistakes.

Keep your hands on the wheel

Teams that get the most out of Cursor write the project's conventions down as rules, pick the right model for each task and read every change the agent makes before it is merged.

Less typing, sharper review

An experienced developer with Cursor writes far less code by hand. The time that frees up goes into design and review, which is where quality is decided.

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.

1

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.

2

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.

3

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.

4

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.

5

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.

1

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.

2

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.

3

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.

4

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.

5

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.

Frequently Asked Questions

What is Cursor?

Cursor is a code editor with AI built in, based on VS Code. It completes your code as you type, answers questions about your project and can run whole tasks with an agent: editing files, running commands and running tests. You choose from models by several providers, such as Claude, GPT, Gemini and Grok.

Is Cursor free?

There is a free version with a limited number of AI requests. For more usage you need a paid plan, from 8 to 200 dollars a month for individuals and per user for teams. Models from other providers, such as Claude and GPT, are billed separately on top of the plan. Current pricing is on cursor.com.

What is the difference between Cursor and VS Code?

Cursor is built on VS Code and looks almost the same. You can bring your settings and most extensions with you. The difference is the AI: completions, a chat about your code and an agent are built into Cursor, and the editor is designed around them.

What is the difference between Cursor and Claude Code?

Cursor is an editor where you type and the AI completes code or carries out a task. Claude Code is an agent you give a task to, which then carries it out in your project from the terminal, your editor or a desktop app. Plenty of developers use both: Cursor for their own writing, Claude Code for handing off bigger tasks.

What is the difference between Cursor and Lovable?

Cursor is built for developers writing code in an existing project. Lovable is built for people who don't write code: you describe an app in a chat and Lovable builds it and puts it online. An app that started in Lovable can later be developed further in Cursor by a developer.

Is Cursor safe for company code?

Cursor sends parts of your code to the AI models to produce an answer. There is a privacy mode in which your code is not stored or used to train models, and the team and business plans give administrators more control. For sensitive projects, check which agreements your organisation needs before you start.

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.

Let's discuss your project

From AI prototypes that need to be production-ready to strategic advice, code audits, or ongoing development support. We're happy to think along about the best approach, no strings attached.

010 Coding Collective free consultation
free

Free Consultation

In 1.5 hours we discuss your project, challenges and goals. Honest advice from senior developers, no sales pitch.

1.5 hours with senior developer(s)
Analysis of your current situation
Written summary afterwards
Concrete next steps