Hey everyone! I hope you are all doing well today. I have been spending a lot of my free time lately diving deep into full-stack web development, and like many of you, I have become pretty reliant on ChatGPT to help me get through the tricky parts of my projects.
Currently, I am working on a React-based dashboard with a fairly complex Node.js and Express backend. It has been a great learning experience, but I have hit a bit of a wall when it comes to efficiency. Right now, my workflow involves a massive amount of alt-tabbing between VS Code and my web browser. I copy a block of code, paste it into the ChatGPT window, ask for a fix or a new feature, and then carefully copy the result back into my editor. It worked fine when the project was just one or two files, but now that I have dozens of components and utility functions, it is becoming a total nightmare to manage. I often lose track of which version of a function I am looking at, and the context switching is killing my productivity.
I am really looking for a way to streamline this process and bring the power of AI directly into my workspace. I have heard people mention a few different options, but I am feeling a bit overwhelmed by the choices and do not want to waste time on tools that do not actually work well. I am specifically looking for tools that offer:
I have looked into GitHub Copilot, but I am curious if there are other extensions or even separate editors, like Cursor, that leverage the ChatGPT API more effectively for a conversational workflow. I really want to feel like the AI understands my whole project structure. Does anyone have a setup they absolutely love or a specific plugin that changed the game for them? What are the best AI tools currently available for coding specifically with ChatGPT or similar models?
Ngl, I disagree about switching editors. Just get Sourcegraph Cody Pro Plan for VS Code. It handles massive project context way better and its more cost-effective for full-stack work tho.
Tbh, you should look into Anysphere Cursor AI Code Editor. Its a fork of VS Code that indexes your whole codebase locally. This lets you use @ symbols to reference specific files or your entire folder as context. If you want to stay in VS Code, Continue.dev VS Code Extension is a solid alternative. Both use RAG to feed your project structure into the model, which totally fixes that annoying context switching nightmare.
Ive been there while building a high-traffic Node app. The turning point was moving to a setup that uses local vector embeddings to index my codebase. It provides the model with the exact context it needs without hitting token limits constantly. Ngl, having the AI understand my directory structure was the productivity win I needed. Definitely look for tools that support full-project indexing tho.
🙌
Commenting to find later
I was looking for something similar a few months back but unfortunately I had a pretty bad experience that made me rethink the whole thing. My buddy was so hyped about moving away from his standard setup to try something more integrated. He thought it was gonna save him hours, but: