Cline + TOKNADO
How to use Cline with TOKNADO's OpenAI-compatible endpoint
Cline is an autonomous coding agent for VS Code. Its OpenAI Compatible provider takes a base URL, an API key and a model ID, which is everything TOKNADO needs.
The setup is quick, but one step is easy to skip and affects quality: telling Cline the model's context window and output limit, because a custom endpoint doesn't always report them.
Start free trial Jump to setup
What you need
- VS Code with the Cline extension installed.
- Your TOKNADO endpoint and API key.
- A model ID from your TOKNADO dashboard.
Setup: Cline with TOKNADO
Open Cline settings
Click the gear icon in the Cline panel. If it's your first launch, choose to use your own API key to reach provider settings.
Choose the provider
Set API Provider to OpenAI Compatible (not “OpenAI”, which is for OpenAI's own API).
Enter the connection details
Use the version-prefixed base URL, with no
/chat/completions. Paste the key with no “Bearer” prefix.Base URL: https://toknado.up.railway.app/v1 API Key : YOUR_TOKNADO_API_KEY Model ID: toknado/autoSet the model limits
Under Model Configuration, set the context window and max output tokens to match the model you're using. Cline can't always discover these from a custom endpoint, and wrong values cause truncation or errors on long tasks.
Click Verify, then test on a throwaway task
Verify confirms the URL, key and model resolve. Then run a small task in a scratch folder before pointing Cline at real code.
Notes for real-world use
- Cline plans, edits across files and runs commands with approval steps, so a single task is many requests. Match your plan to how often you run it.
- If Verify fails, the cause is almost always the base URL, the key or the model ID, in that order. A direct curl test isolates which.
- You can keep several providers configured and switch between them for different jobs.
Troubleshooting Cline + TOKNADO
| Symptom | What to check |
|---|---|
| Verify fails immediately | Check the base URL ends at the API root (usually /v1) and the key has no extra characters. |
| “Model not found” | Enter the model ID exactly as shown in your TOKNADO dashboard. |
| Connection errors | Make sure the endpoint is reachable from your machine and no firewall or proxy blocks it. |
| Long tasks truncate or error | Set realistic context window and max output values under Model Configuration. |
| Works in curl, not in Cline | Confirm you selected OpenAI Compatible, not OpenAI. |
FAQ
How do I add a custom provider to Cline?
Open settings, set API Provider to OpenAI Compatible, then enter the Base URL, API key and Model ID.
Should the base URL include /chat/completions?
No. Use the /v1 style base; Cline adds the path.
Why set the context window manually?
Custom endpoints often don't report model limits, and Cline needs them to manage long conversations safely.
Does the trial work with Cline?
Yes. The TOKNADO trial gives up to 250 requests per rolling hour for three days, enough to confirm your setup.