Just tried a direct test. I gave both Cursor (with Composer) and ChatGPT (4o) the exact same prompt: "Generate a set of Node.js/Express CRUD endpoints for a 'tasks' model with MongoDB/Mongoose."
Here's what I noticed:
Cursor felt faster for the actual editing. It wrote the code directly into my open project files, which was huge. The context of my existing code seemed to help it structure things more cleanly. It auto-imported the required modules correctly.
ChatGPT gave a solid, generic answer in the chat, but then I had to copy, paste, and adjust everything manually. It didn't know about my project structure, so I spent extra time fixing import paths.
For a bootstrapped founder like me, that workflow difference is a big deal. Saving 10 minutes on boilerplate setup matters. Cursor's free tier handled this perfectly. Anyone else done a similar side-by-side? Curious about your results, especially for more complex endpoints.
Free tier is my favorite tier.