Trae Setup

Installation

Download the installer for your platform (Windows / macOS / Linux) from the Trae website, then run it — no Node.js or npm required.

Configure a Custom Model

1. Open Settings → Models to access the model management panel.

2. Click the Add Model button to open the add model dialog.

3. Select the Custom Configuration option.

4. Set the API format to OpenAI Chat Completions format.

5. Enter the request URL: turn off the Full URL toggle, and enter the base address:

   https://tokenfly.com/v1

6. Enter the model ID:

   gpt-5.4

7. Enter the API key: sk-xxx.

8. Click Add Model.

Trae will call the provider's API to verify the API key: on success the model appears in the model list; on failure an error message and provider log are shown to help troubleshoot.

Configuration Reference

Switching Models

Click the current model name in the bottom-right corner of the chat input box to open the model list, then select gpt-5.4.

Verification

After selecting the gpt-5.4 model, start a conversation — a normal reply confirms the setup works.