Configuring Zcode
Zcode is Zhipu AI's coding client. In addition to its built-in BigModel service, it supports OpenAI-compatible custom providers.
1. Add the wylon provider
- Select the gear icon.
- Go to Basics → Model settings.
- Under Custom providers, select + Add provider.
Provider information
| Field | Value |
|---|---|
| Name | wylon-cloud, or another display name |
| Base URL | https://api.wylon.cn/v1 |
| API key | Your wylon API key |
| API format | Chat completions (/chat/completions) |
Base URL and API format
Enter the base URL through /v1; Zcode appends the request path.
warning
Do not append the request pathDo not put
/chat/completions in the Base URL. Zcode appends it automatically, and a duplicated path returns 404.2. Add a model
Under Model list, choose + Add model and enter:
| Field | Example |
|---|---|
| Model ID | ZhipuAI/GLM-5.3 |
| Context window | 1000000 |
| Max output tokens | 128000 |
| Input types | Text only for this example |
| Output types | Text |
Model ID
Use the exact, case-sensitive ID from the Model Plaza or GET https://api.wylon.cn/v1/models.
warning
The model list is not an access guarantee
GET /v1/models may include a retired model or one that the current API key cannot access. Verify it with a minimal text request before adding it; the test request incurs normal model usage.Input types
Match input types to the model's actual capabilities. Then return to the provider form and select Add provider.
3. Verify the configuration
- Confirm that
wylon-cloudappears under Custom providers and is enabled. - Select its model in chat.
- Send a simple message and confirm that it returns normally.
4. Add or switch models
Copy a complete ID from the Model Plaza or GET /v1/models. Keep the provider URL and key, then add a model or replace an existing model ID.
