Choose models and endpoints
Start in the model marketplace
Open the marketplace, choose your key's group, and check the model name, price, availability and supported interfaces.
Use the exact model ID, not a custom display name. Preserve capitalization, suffixes and dates.
Fetch the model list
In a Bash terminal with GUANGJI_API_KEY set:
curl 'https://guangjiapi.top/v1/models' \
-H "Authorization: Bearer $GUANGJI_API_KEY"
The result typically includes a data array with model id values. Visibility depends on permissions and configuration. Listing a model does not guarantee every parameter or request will work at all times.
Match the protocol
| Use case | Common endpoint |
|---|---|
| OpenAI-compatible chat | POST /v1/chat/completions |
| Responses | POST /v1/responses |
| Claude Messages | POST /v1/messages |
| Image generation | POST /v1/images/generations |
| Embeddings | POST /v1/embeddings |
Supported endpoints and capabilities vary by model. Check the integration methods currently available through Guangji API before choosing a request format.
Model not found or unavailable
- Copy the model ID again from the marketplace.
- Check the key's group and model restrictions.
- Check the provider, protocol and endpoint configured in the client.
- Review announcements or retry later. If it persists, send the request ID to support.
Avoid switching to an unverified model without checking its price and purpose.