Suanlimao
← Back to all tutorials

Connect your tools

5ire

The desktop AI assistant with MCP support: add an OpenAI-compatible provider.

About 4 minFor:People who chat and use MCP tools in 5ire

Before you start

  • 5ire is installed
  • A key and a model name

Step 1 · STEP 0x01

Add an OpenAI-compatible provider

Providers
1
Add
2

Type

OpenAI Compatible
3

Name

Anything you like
Diagram only — the real screen in your version may differ

Add a new provider and choose OpenAI Compatible. 5ire supports custom OpenAI-compatible providers since v0.10.0 and lets you switch between them mid-conversation.

Create a key →

Step 2 · STEP 0x02

Enter the endpoint and key

OpenAI Compatible → Edit
1

Endpoint

https://api.suanlimao.shop/v1
2

API Key

<YOUR_API_KEY>
3

Proxy

Leave empty
4
Save
Diagram only — the real screen in your version may differ
  • Enter https://api.suanlimao.shop/v1 as the endpoint; the docs do not say whether /v1 is needed, so on 404 try https://api.suanlimao.shop.
  • Proxy is for routing 5ire through a network proxy, not our address — leave it empty.
5ire has no formal user manual; the labels above come from the app's own Chinese strings, so follow your screen.

Open the model plaza →

Step 3 · STEP 0x03

Add a model and chat

OpenAI Compatible → Models
1
Add model
2

Model name

<MODEL_NAME>
3

Model picker in chat

<MODEL_NAME>
Diagram only — the real screen in your version may differ
  • Add a model under the provider (name copied from the plaza), pick it in chat and send a message.
Don't guess the model name: open the model plaza, click “Copy model id” on a model card and paste it. A wrong model name is the most common error.

✓ How to tell it worked

  • You get a reply and the Usage page shows the request.

FAQ

“API endpoint must be a valid URL”?

Use a complete address starting with https://.

404?

The release notes specifically say to check the API Base on 404. Switch between https://api.suanlimao.shop/v1 and https://api.suanlimao.shop, then check the model name.

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).