1. Go to the official website and register an account (must read)

image

image

1. After registration, use a redemption code to redeem quota (you can also top up online here for renewal)

image

2. Then create a Key; you only need to fill in the name under the API Key, adding a group is optional

image

image

3. Copy the created token key

image

4. Now you can configure it in your desired client, e.g. Lobster, CC, Codex, etc. (when configuring, select the OPENAI protocol for custom model configuration)

OpenAI protocol endpoint URL: https://www.yundare.com/v1

Anthropic protocol endpoint URL: https://www.yundare.com

You only need three things: endpoint URL + token + the model name to use (check the 'Model Square' at https://www.yundare.com/pricing to see available models, just copy their names)

Can't connect? How to quickly verify?

In the Playground at https://www.yundare.com/playground, select a model. If the model responds normally, the configuration in other clients is wrong; otherwise it is a model-relay issue. You can switch to another model or contact us for feedback

image

2. Claude Code Configuration Method

1. First download the latest version of CC Switch (optional, but more convenient). CC Switch download URL:

[CC-Switch-v3.17.0-Windows.msi]

2. Go to https://www.yundare.com/keys for API Key management

image

Select / fill in the relevant connection information

image

Import the configuration in CC Switch

image

3. After importing, click Test to see the status

image

4. Open the terminal

image

5. The Claude Code graphical plugin on VSCode can also be used (remember to restart VSCode)

image

6. Configuration Issues

Issue 1

(If you encounter this problem, it is because the Claude plugin does not support the OpenAI protocol; refer to Figure 2 to select and set the OpenAI protocol and enable routing)

image

Change the API format to the OpenAI protocol

image

Enable routing settings and open Claude with routing enabled

image

Issue 2

image

1. Update to the latest version.

image

2. Switch to the Codex native protocol (no routing needed)

image

3. Exit Codex completely, end the process and quit, then reopen

3. Codex Configuration Method

1. First download CC Switch (optional, but more convenient). CC Switch download URL:

Codex download URL:

2. Go to https://www.yundare.com/keys for API Key management

image

Select / fill in the relevant connection information

image

Import the configuration in CC Switch

image

4. After importing, click Test to see the status

image

5. Restart Codex to start chatting. [If you did not configure a GPT-series model, the model field will be blank, but it is still usable.] [When blank, it calls the first model you selected]

[CC needs to restart Codex every time you switch models]

image

3. Using in Roo Code / Kilo Code

API provider: select Anthropic

Fill in the API Key

Use a custom base URL, fill in https://www.yundare.com/

image

4. Using in Cherry Studio

[Cherry Studio.exe]

Add a provider platform

Type: select Anthropic

Fill in the API Key; API address: https://www.yundare.com/

Click Manage to pull models and add them

image

image

5. OpenClaw Model Configuration

You need 3 things. Base URL: https://www.yundare.com/v1

API Key: (the key obtained in the Key creation step) Model ID: (the model name under the selected group)

JSON


Method1:#
Edit the configuration file~/.
openclaw/openclaw.json#
FindbaseUrl, apiKey, Model ID replace then save{
 "models": {
 "providers": {
 "
aiii": { "
baseUrl": "https://www.yundare.com/v1", "
apiKey": "sk-yourAPIAPI Key", "
api": "openai-completions", "models": [
 {
 "id": "gpt-5.6-luna",
 "name": "gpt-5.6-luna",
 "
contextWindow": 1048576 }
 ]
 }
 }
 }
}


# Restart the gateway to apply the configuration
openclaw gateway restart

# Check whether the model was added successfully
openclaw model list

Codex relay beginner's tutorial!: low-price tokens at https://www.yundare.com/