04/03/2026 1:31 am
Topic starter
Notifications
Clear all
Other LLMs Forum
3
Posts
4
Users
0
Reactions
118
Views
0
So I've been hitting a wall with this data automation script I'm writing for a freelance client. It's all Python, mostly pandas and some weird API stuff. I need to get it done by next Friday or I don't get paid lol so I'm trying to pick a paid tool to speed me up.
I'm totally torn between just getting a GitHub Copilot subscription or switching my whole IDE over to Cursor. My logic was that Copilot is just easy because it's right there in VS Code and I'm already used to the workflow, but then I see everyone online saying Cursor is basically a Copilot killer because it indexes the whole project better. I'm also considering just sticking with ChatGPT Plus and copy-pasting but honestly that's getting really annoying with larger files and I keep losing track of which version of the code is the right one.
I've got a strict $20 a month budget for tools so I can really only pick one right now. I'm leaning toward Cursor but I'm worried about the learning curve when I'm on such a tight deadline. Has anyone here used both for heavy Python work? Is the codebase indexing in Cursor actually worth the switch or should I just stick with what I know in VS Code and get Copilot...
3 Answers
12
04/03/2026 1:31 am
Honestly, I had issues with Cursor and it felt way too clunky for twenty bucks. Unfortunately, the indexing isnt as reliable as people claim and it messed up my pandas imports. I would play it safe to protect your budget.
- Get GitHub Copilot Individual Plan
- Stick with your VS Code setup Learning a new tool mid-project is just too risky when you need that paycheck.
10
04/03/2026 2:35 am
Unfortunately, I've had issues with those flat-rate subscriptions. They often throttle your speed right when you're on a deadline, which isnt as good as expected for twenty bucks. To save money, Id honestly suggest Anthropic Claude 3.5 Sonnet API via a pay-as-you-go extension. Its way more budget-friendly because you only pay for what you use, and the Python reasoning is better for pandas... worth a look.
2
04/03/2026 1:55 am
To add to the point above: I actually tried using Sourcegraph Cody for a similar data project. Unfortunately, I had issues with the repository indexing. It wasnt as good as expected. The AI frequently hallucinated library functions that dont exist. I found the manual sync process quite inefficient. Technical errors occurred frequently. It was honestly a disappointing experience for a paid tool.
