Applicable Scenarios
Cline is an AI coding assistant for VS Code. It supports an OpenAI Compatible provider, so you can directly configure Tokener CA as a custom model provider for code understanding, editing, command execution confirmation, and project-level Agent tasks.
For the integration you need to prepare:
| Configuration Item | Value |
|---|---|
| API Provider | OpenAI Compatible |
| Base URL | https://a.tokener.ca/v1 |
| API Key | API key, for example tok-... |
| Model ID | Platform model name, for example glm-5.3:flash |
Configure in Cline Settings
Open the Cline panel in VS Code, go to the settings page, and fill it in as follows:
| Field | What to Fill In |
|---|---|
| API Provider | Select OpenAI Compatible |
| Base URL | https://a.tokener.ca/v1 |
| API Key | API key |
| Model | glm-5.3:flash, or the platform model ID you want to use |
Note: Cline's Base URL should stop at /v1; do not enter the full /v1/chat/completions. Cline appends the chat path automatically, following the OpenAI-compatible protocol.
Model Capability Configuration
Cline lets you specify the context length, maximum output, image capability, and tool capability for a custom model. We recommend starting with conservative values and then adjusting them based on the model page:
| Capability Item | Recommendation |
|---|---|
| Context Window | Fill in according to the model page's context length; when unsure, first fill in 32768 or 128000 |
| Max Output Tokens | First fill in 4096 or 8192 |
| Image Support | Only enable when the target model supports image input |
| Computer Use / Tool Use | Enable when you need Cline to perform tool or code-editing tasks |
| Input / Output Price | Can be filled in according to the platform model page prices, used for local cost estimation |
Verify the Connection
After saving the configuration, send a small task in the Cline conversation:
Please read the currently open file and explain what it does in three sentences.If Cline returns a result and the corresponding request appears in the Tokener CA console's activity records, the integration is working.
You can also get curl working first with the same key and model:
curl -X POST "https://a.tokener.ca/v1/chat/completions" \ -H "Authorization: Bearer <consumer_api_key>" \ -H "Content-Type: application/json" \ --data-raw '{ "model": "glm-5.3:flash", "messages": [ { "role": "user", "content": "Please return a short message confirming the integration succeeded." } ], "max_tokens": 64, "stream": false }'Troubleshooting Checklist
| Symptom | How to Handle |
|---|---|
| Authentication failure | Check whether the API Key was copied completely and whether it is filled in the OpenAI Compatible API Key field |
| Model does not exist | Check whether the Model ID exactly matches the platform model list |
404 | Check whether the Base URL is https://a.tokener.ca/v1, without adding /chat/completions |
| Output is very short | Increase Max Output Tokens |
| Tool calls unstable | Confirm the model is suitable for tool calling and enable the corresponding capability in the Cline model configuration |