How to Add OpenRouter.ai to Cursor

So you're ready to unlock custom models in Cursor, eh? Good. Here's exactly how to plug in OpenRouter.ai so you can run stuff like z-ai/glm-4.5 directly inside your coding flow.

Follow these steps carefully, or you’ll just end up confused and yelling at your terminal.

🚀 Step-by-Step: Integrating OpenRouter.ai into Cursor

Open Cursor Settings
Click the gear icon (bottom-left). Yes, the thing you’ve ignored until now.
Navigate to Models
Once inside settings, go to the Models section.
Disable ALL Current Models
That’s right. Every single one. Disable anything that’s checked or active.
Scroll Down to > Api Keys
You’ll find a collapsible section labeled > Api Keys. Click it to expand.
Override the OpenAI Base URL
Under the OpenAI API Key area:
- Check the box that says Override OpenAI Base URL
- Paste this into the field:
  OpenRouter base URL
```
https://openrouter.ai/api/v1
```
Find the Right Model Name
This part matters. Go to https://openrouter.ai/models.
Find the proper name of the model you want to use, for example z-ai/glm-4.5. You’ll need to type it exactly like that when adding a new model.
Add the Model in Cursor
Back in Cursor:
- Click Add New Model
- Paste the model name from OpenRouter (e.g., z-ai/glm-4.5)
- Save it
Add Your OpenRouter API Key
Paste your API key into the OpenAI API Key section.
Then hit Verify.

If everything goes well, you’ll see something like:

With your cursor Pro+ subscription, you do not need to use your own OpenAI key unless you're using custom models.
Test the Integration
Open the Cursor chat and try talking to your newly added model.
If it replies, congrats, you’re golden.

✅ Bonus Tips

You can add multiple models from OpenRouter.ai now.
Check out z-ai/glm-4.5-air:free: it’s one of the free models and surprisingly solid.

⚠️ Heads up: Free-tier models may use your prompts and completions as training data. So… maybe don’t use them for your nuclear launch codes.

Happy hacking.