Cursor

The 80/20 verdict

Cursor is the default AI code editor for developers who write code every day and want LLM assistance that’s actually integrated into their workflow, not bolted on. GitHub Copilot is the incumbent and it’s fine, but it’s a plugin — Cursor is built from the ground up around the model, which means the context window, the chat sidebar, and the multi-file editing are all first-class features rather than afterthoughts. If you’re already paying for Copilot and not feeling like it’s changing how you work, try Cursor for a week. It’s that kind of difference.

What is Cursor

Cursor is an AI-native code editor built on a VS Code fork. It replaces GitHub Copilot not just as a completion tool but as an active editing partner: you can highlight code and describe a change in plain English (Cmd+K), ask questions about the entire codebase in a chat sidebar, or use Composer to make coordinated changes across multiple files at once. Because it’s a VS Code fork, all your existing extensions, themes, and keybindings carry over with zero configuration. The AI layer runs on Claude or GPT-4o and can use your own API keys if you hit the monthly request cap.

Key features

• Tab completions that predict multi-line changes based on surrounding context, not just the line above
• Cmd+K inline editing — describe a change in plain English and apply it to a selection
• Chat sidebar with full codebase indexing for explaining, debugging, and refactoring
• Composer mode for multi-file changes across the entire project in one session
• Bring-your-own API key support to bypass monthly request quotas
• Built-in terminal with AI assistance for interpreting and fixing shell errors
• VS Code extension compatibility — existing extensions and themes work without reinstalling

When to use it

• You write code daily and want completions that understand the full file, not just the line above the cursor. Cursor’s tab completions predict multi-line changes and entire function bodies based on what you’ve written nearby.
• You’re refactoring or building a new feature and want to chat with the codebase. Cursor’s Cmd+K and chat features let you ask questions about your own code and get edits applied directly to the file.
• Your team is already paying for GitHub Copilot at $19/user/mo. Cursor Pro at $20/mo replaces it with a better model, better context, and a better interface.

When to skip

• You’re happy with Vim or Emacs and don’t want to move to a VS Code fork. Cursor is VS Code under the hood — it’s excellent, but if your muscle memory is in a different editor, the switch has a cost.
• You work in a regulated codebase where no code can be sent to external APIs. Cursor’s completions send code context to Anthropic or OpenAI. Use a locally-hosted model with Continue.dev instead.
• You’re a hobbyist who codes a few hours a week. The free tier’s completions will run out fast on daily use; $20/mo is only worth it if coding is your job.

Pros and cons

Pros:

• Context window is genuinely larger than Copilot — understands whole files and imports, not just nearby lines
• Multi-file Composer mode changes code across multiple files in one coordinated session
• VS Code-compatible — zero migration cost for the majority of working developers
• Model-agnostic: use Claude, GPT-4o, or your own API key — not locked to one provider
• Chat explains what it changed and why, making it educational as well as generative

Cons:

• Sends code context to Anthropic or OpenAI — a hard blocker for regulated or proprietary codebases
• 500 premium requests/mo on Pro burns fast on large codebase queries; bring-your-own-key helps
• $20/mo requires daily coding use to justify the cost
• Still VS Code under the hood — Vim and JetBrains users face a real migration

Who is using Cursor

Cursor has become the default editor for a significant portion of YC-batch founders who write their own code — it’s particularly common among technical co-founders building fast and alone. AI engineers, full-stack developers, and anyone whose job is primarily writing and refactoring code are the core users. Freelance developers use it to work faster on client projects. Teams that adopted it early report it’s now the first tool they miss when switching machines. It’s less common in large enterprise engineering orgs where regulated codebases and Jira-driven workflows create friction with AI code tools.

Pricing reality check

The Free tier gives 2,000 code completions and 50 “slow” premium model requests per month — enough to evaluate but not enough for daily professional use. Pro at $20/mo gives unlimited completions and 500 fast premium requests monthly, plus the ability to use your own API keys to bypass the cap. Business at $40/user/mo adds centralized billing and usage policies. The pricing is straightforward; the only surprise is that premium model requests (Claude Sonnet, GPT-4o) burn faster than expected on large codebase queries — plan on using bring-your-own-key if you’re doing heavy refactoring work.

What makes Cursor unique

Cursor was built ground-up for LLM integration rather than bolted on as a plugin. The context window, multi-file editing via Composer, and chat interface are first-class product features, which makes the AI assistance materially better than Copilot’s plugin architecture. No other editor in the category has the same depth of integration between the editing surface and the model.

How we evaluated Cursor

We last verified Cursor’s pricing and features on 2026-05-24 by reviewing the pricing page and the product changelog. See our evaluation methodology.