Suanlimao
← Back to all tutorials

Connect your tools

Cursor

Point Cursor at our OpenAI-compatible address and add models by their alias.

About 6 minFor:Developers who code in Cursor

Before you start

  • Cursor is installed and you are signed in
  • A key and a model name copied from the model plaza
  • Read the warnings below first — Cursor has its own limits

Step 1 · STEP 0x01

Open the Models page in Cursor Settings

Cursor Settings → Models → API Keys
1

OpenAI API Key

<YOUR_API_KEY>
2

Override OpenAI Base URL

On
3

Base URL

https://api.suanlimao.shop/v1
4
Verify
Diagram only — the real screen in your version may differ

Go to Cursor Settings → Models and scroll to the API Keys section. Under OpenAI API Key paste your key, turn on Override OpenAI Base URL, enter the address in the field that appears and click Verify.

Use https://api.suanlimao.shop/v1 as the API address (with /v1). If the tool says 404, check whether /v1 is missing or doubled, then read the FAQ below.
Every request from Cursor goes to Cursor's own servers first and is forwarded from there, so the source IP is not your computer. The address must be reachable from the internet — never localhost or an internal address — and any IP restrictions on your key should be read with that in mind.

Create a key →

Step 2 · STEP 0x02

Add the model by its alias

Cursor Settings → Models
1

Add model

<MODEL_NAME>
2
Add model
3

Model switch

On
Diagram only — the real screen in your version may differ

With a custom address Cursor's own model list no longer applies, so add the model names you want by hand. Write the name as <MODEL_NAME>: the model id copied from the plaza with the prefix in front. The model details drawer has a ready-made “Copy alias” button.

Cursor rejects custom model names that start with claude- or gemini- (ERROR_BAD_MODEL_NAME, or it treats the request as one of its own models and ignores your key). The prefixed alias is exactly the same model — same service, same price — and the Usage page records it under the original model id.

Open the model plaza →

Step 3 · STEP 0x03

Try a message in chat

Cursor Chat
1

Model picker

<MODEL_NAME>
2

Message box

Hello
3
Send
Diagram only — the real screen in your version may differ

Pick the model you just added and send a message. A normal reply plus a new row on the Usage page means it works.

Versions behave differently: some apply the override to chat only, with Agent and Tab still using Cursor's own models, and some make Cursor's built-in models unavailable while the override is on. Trust what your version shows.

✓ How to tell it worked

  • Chat answers with the model you added.
  • The Usage page shows the request under the original model id without the prefix, proving it went through our service.

FAQ

I cannot find Override OpenAI Base URL.

It sits in the API Keys section of the Models page, right under OpenAI API Key; some versions only show it after a key is entered. If your version has no such option use Cline or Continue instead; see their tutorials.

I get ERROR_BAD_MODEL_NAME.

You entered the plain model id. Add it again as <MODEL_NAME>.

Connection failed / cannot reach the address.

Requests leave from Cursor's servers, so the address must be reachable from the internet. Do not use localhost, an internal address or an office-only gateway.

Model does not exist?

The part after the prefix must match the plaza exactly. Copy it again without extra spaces.

Can Agent mode use our model?

It depends on your Cursor version and plan and the official docs do not say. If it does not work use Cline or Claude Code instead.

Related tutorials

Still stuck?

Read the troubleshooting tutorials first. If that does not help, open a ticket with the request_id, time, model and the exact error (never paste your key).