DeepSeek with Cursor: A Hack to Use Your Own API Key

If you're like me and spend hours and hours a day inside Cursor, you probably appreciate how smooth the dev experience is, especially with OpenAI models.

But when I tried every model Cursor supports by default, I started wanting more. I had some leftover credits on DeepSeek, and figured, “Why not try it in Cursor?”

But wait! Seems like I can’t do that 🤔

Cursor supports custom API keys for OpenAI, Anthropic, Google, Azure, but not for DeepSeek.
So even though DeepSeek has amazing models like deepseek-reasoning, there's no native way to plug in your own API key.

This is especially frustrating if:

• You're on a free Cursor plan or used monthly limits of Pro.

• You already have credits on DeepSeek

• You want to try the latest models without waiting for Cursor to officially support them

So… I hacked it in.

Let’s walk through how I got my personal DeepSeek account in Cursor using my own API key 🔧

🧠 Step-by-step: How to Use DeepSeek API in Cursor

Step 1: Get an API key 🔑 from DeepSeek platform

Step 2: Go to "Cursor Settings" and select "Models" below the settings

Paste your API key in OpenAI API Key field and override OpenAI Base URL with this one https://api.deepseek.com/v1

Create 2 new models by clicking + Add model on the same page. Models should be deepseek-chat and deepseek-reasoning 

Lastly, uncheck all other models and leave only deepseek-chat and deepseek-reasoning selected. Then, go to the OpenAI API section and click 'Verify'. You should see a success message and that’s it, you're all set 🎉

Let me know if this kind of content is valuable to you, I’m here to help you level up and become a truly 10x AI-powered engineer. My goal is to share practical insights, tools, and workflows that actually make a difference in your day-to-day.

Talk soon,
George