AI-narrated version of this post using a synthetic voice. Great for accessibility or listening while busy.
After six months of daily use, our honest read on Cursor is that it earns its reputation in some areas and quietly underdelivers in others. Fullstack and Python developers who spend most of their time on multi-file refactors will likely find Cursor’s Composer feature alone makes the subscription defensible – casual coders or teams working primarily outside JavaScript ecosystems will have a harder time justifying the cost. This review covers what we actually found in regular use: where it holds up, where it frustrates, and whether it makes sense to pull away from your current editor setup.
What Is Cursor?
Cursor is an AI-first code editor built as a fork of Visual Studio Code. It’s developed by Anysphere and ships with deep AI integration baked directly into the editor â not bolted on as an extension. That’s the core architectural difference between Cursor and something like VS Code + GitHub Copilot: the AI features are first-class citizens here, not afterthoughts.
Because it’s a VS Code fork, you get the full extension marketplace, familiar keyboard shortcuts, and the same general feel you’re probably already comfortable with. Cursor layers on top of that with a chat panel, inline edits, multi-file Composer mode, and the ability to switch between models including Claude Sonnet, Claude Opus, GPT-4o, and Gemini. It runs on Mac, Windows, and Linux. The company has been iterating quickly through 2025 and into 2026, with agent mode becoming increasingly central to the product’s identity.
What Cursor Does Well
Composer is genuinely the best multi-file editing experience I’ve used. This is where Cursor earns its keep. When you need to refactor a feature across eight files, update an API contract, and adjust the corresponding tests all in one go, Composer handles it with a coherence that VS Code + Copilot simply doesn’t match. You describe what you want, Cursor shows you a diff across every affected file, and you can accept or reject changes file by file. It’s methodical and transparent.
Agent mode extends this further by letting Cursor run terminal commands, read error output, and iterate autonomously. I’ve used it to scaffold entire Next.js features, run migrations, and fix cascading TypeScript errors without babysitting every step. It’s not magic â it still needs guidance â but it saves real time on the kind of tedious multi-step work that used to require significant manual coordination.
Model switching is practical, not just a feature checkbox. Being able to toggle between Claude Sonnet for fast iteration, Claude Opus for complex reasoning tasks, and GPT-4o when you want a different perspective is genuinely useful. Different models have different strengths on different codebases, and having them accessible from one tool without juggling browser tabs is a legitimate quality-of-life improvement.
The VS Code foundation means your existing workflow survives the switch. Your extensions come with you. Your themes, your keybindings, your git integration â all of it transfers. The migration friction is low, which matters when you’re evaluating whether a tool is worth disrupting your setup.
Context handling has also improved noticeably. Cursor’s @codebase and @file references let you point the AI at specific parts of your project, and the retrieval is accurate enough that I rarely find it hallucinating about my own code structure. That wasn’t always true in earlier versions.
What Cursor Does Poorly
It lags behind VS Code mainline, and that matters. Because Cursor is a fork, it’s always chasing VS Code’s update cycle. In practice, that means you’re sometimes running an older version of the underlying editor while Microsoft ships improvements. For most users this is a minor annoyance, but if you depend on specific VS Code features that ship in point releases, you may find yourself waiting weeks for Cursor to catch up. It’s an inherent structural limitation of the fork model.
Pricing can balloon quickly on heavy use. The Pro plan runs $20 USD ($27 CAD) per month, which sounds reasonable until you hit the usage limits on premium model requests. Cursor uses a credit-style system for the more capable models, and if you’re running Composer sessions with Claude Opus across large codebases daily, you’ll chew through those credits. Upgrading to Business at $40 USD per user per month helps, but teams doing serious AI-assisted development should model their actual usage before committing. Compare that to GitHub Copilot at $10-$19 USD per month, and the cost difference requires justification.
UI bugs are more common than they should be at this price point. In my time using Cursor, I encountered a persistent issue where the chat panel would lose context mid-conversation and require a restart. I also hit occasional rendering glitches in the diff view during Composer sessions. These aren’t showstopping bugs, but they’re the kind of thing that chips away at trust in a tool you’re paying for monthly. The team ships fixes regularly, but the bug rate feels higher than a mature product warrants.
Non-Node environments get less love. If your stack is primarily Python data science, Rust systems programming, or anything outside the JavaScript/TypeScript ecosystem, Cursor is fine but not exceptional. The AI suggestions, the Composer workflows, and the documentation all feel tilted toward web development. Python support has improved but still trails the JS experience meaningfully.
Pricing (2026)
Cursor’s current pricing breaks down as follows:
- Hobby (Free): 2,000 completions, 50 slow premium requests per month. Enough to evaluate, not enough to rely on professionally.
- Pro: $20 USD / approximately $27 CAD per month. 500 fast premium requests, unlimited slow requests, access to all models.
- Business: $40 USD / approximately $54 CAD per user per month. Adds team management, SSO, centralised billing, and privacy mode by default.
The practical ceiling on Pro is real. If you’re doing several hours of Composer work daily with Claude Opus or GPT-4o, you’ll likely hit the fast request limit and either slow down or look at Business. Budget accordingly.
How It Compares
VS Code + Copilot is the obvious comparison. Copilot is cheaper and deeply integrated into Microsoft’s ecosystem. For inline completions and simple chat, the gap has narrowed significantly. But Copilot’s multi-file editing still doesn’t match Composer’s coherence, and the agent mode is less mature. If multi-file refactoring is central to your work, Cursor wins that specific comparison clearly.
Windsurf from Codeium is the closest direct competitor. It also offers agentic editing and is similarly priced. From my testing, Windsurf’s Cascade agent is competitive with Cursor’s Composer, and some developers find its UI less buggy. It’s worth a trial if you’re evaluating the category rather than committed to Cursor specifically.
Claude Code is a different animal â a terminal-first agentic tool rather than a full editor. It’s more powerful for autonomous multi-step tasks but lacks the editor experience entirely. Some teams use both: Cursor for daily editing, Claude Code for larger autonomous tasks.
Who Should Buy Cursor
Cursor is a strong fit if you’re a fullstack web developer or TypeScript-heavy engineer who regularly needs to make coordinated changes across large numbers of files. It’s also well-suited for Python developers working on application code (as opposed to data science workflows). If your team is already on VS Code and you want to upgrade the AI layer without rebuilding your entire workflow, the migration path is genuinely low-friction. Development teams that can justify $40 per seat per month and want centralised control over AI tool usage will find the Business tier practical.
Who Should Skip Cursor
If you primarily write Python for data science, machine learning model training, or notebook-heavy work, the ROI is weak. Jupyter integration is serviceable but not a strength. If you’re budget-conscious and your AI needs are covered by inline completion and occasional chat, VS Code + Copilot at half the price is a more rational choice. If you work in a primarily Rust, Go, or C++ environment, the AI assistance quality drops enough that the premium is harder to justify. Teams with strict data governance requirements should read the privacy documentation carefully before adopting any cloud AI coding tool.
FAQ
Does Cursor send my code to external servers?
Yes, by default. Cursor sends code context to AI providers to generate responses. Business tier includes privacy mode, which disables training data use. Review their privacy policy before adopting it in environments with sensitive intellectual property.
Can I use my own API keys?
Yes. Cursor supports bring-your-own-key for OpenAI, Anthropic, and other providers. This can reduce costs if you have existing API credits, though you lose the bundled request allocation.
Is Cursor worth it for solo developers?
If you’re doing serious daily development work, Pro at $20 USD per month is defensible. If you code occasionally or on side projects, the free tier or VS Code + Copilot is probably sufficient.
How often does Cursor update compared to VS Code?
Cursor typically lags VS Code by weeks to a couple of months on any given update cycle. For most developers this is acceptable. If you need to be on VS Code mainline immediately for any reason, the fork model is a structural constraint you won’t escape.
Verdict
Cursor earns a genuine recommendation for fullstack and TypeScript developers who spend meaningful time on multi-file refactoring work. Composer is the best implementation of that feature category I’ve tested, agent mode is practically useful rather than just demonstrable, and the VS Code foundation keeps the transition cost low. The weaknesses are real though: the pricing structure punishes heavy use, the UI bugs undercut the polished experience the product is clearly aiming for, and developers outside the JavaScript ecosystem will find diminishing returns. At $20 USD per month, it’s worth a trial month to see whether Composer changes your workflow the way it changed mine. If it doesn’t click within a few weeks of real use, VS Code + Copilot or Windsurf are both credible alternatives worth your time.
Try Cursor’s free tier here before committing to a paid plan.
AIToolPickr shares honest AI tool reviews. Some links may earn a small commission at no cost to you. Editorial, not sponsored.
Related Auburn AI Products
Building content or automations around AI? Auburn AI has production-tested kits:
- 100 Claude Prompts for Canadian SMB Owners ($17)
- The n8n + Claude Blog Automation Stack ($47)
- Auburn AI Monitoring Stack ($37)
- Browse the full catalogue
— Auburn AI editorial, Calgary AB