Open Code

Open Code can use HaiToken through an OpenAI-compatible provider. The stable path is a two-part flow: use /connect to store the credential, then use opencode.json to define the provider, Base URL, and models.

Protocol Note

Open Code’s custom provider integration is based on @ai-sdk/openai-compatible, so it only supports the OpenAI Chat Completions protocol; it does not support the Anthropic Messages protocol. This guide therefore only covers the OpenAI mode, with the Base URL fixed to https://api.haitoken.ai/v1.

What you need

A HaiToken API key
At least one available model ID, such as gpt-5.4
HaiToken OpenAI-compatible base URL: https://api.haitoken.ai/v1

Setup Steps

Step 1: Save the HaiToken API key with `/connect`

Inside Open Code, run /connect and store your HaiToken API key there. This step is mainly for credentials. Do not expect a Base URL input in this dialog.

Step 2: Edit `opencode.json`

Use your Open Code config file to define the provider, Base URL, and models. A common global path is ~/.config/opencode/opencode.json. You can read that path like this on different systems:

Windows: a common equivalent is %USERPROFILE%\\.config\\opencode\\opencode.json
macOS: commonly ~/.config/opencode/opencode.json
Linux: commonly ~/.config/opencode/opencode.json

Here ~ means your user home directory: usually %USERPROFILE% on Windows, /Users/<your-name> on macOS, and /home/<your-name> on Linux. Example:

{
  "$schema": "https://opencode.ai/config.json",
  "provider": {
    "haitoken": {
      "name": "HaiToken",
      "npm": "@ai-sdk/openai-compatible",
      "options": {
        "baseURL": "https://api.haitoken.ai/v1",
        "apiKey": "{env:HAITOKEN_API_KEY}"
      },
      "models": {
        "gpt-5.4": {
          "name": "gpt-5.4"
        },
        "gpt-5.4": {
          "name": "gpt-5.4"
        }
      }
    }
  }
}

Step 3: Make sure the environment variable exists

Windows PowerShell:

$env:HAITOKEN_API_KEY = "YOUR_API_KEY"

macOS / Linux:

export HAITOKEN_API_KEY="YOUR_API_KEY"

Step 4: Restart Open Code and choose the model

After saving opencode.json, restart Open Code if the provider list does not refresh automatically. Then select the HaiToken provider and switch to the model you registered.

Request Mapping

Item	Value
Base URL in config	`https://api.haitoken.ai/v1`
Chat endpoint used by Open Code	`POST /v1/chat/completions`
Model discovery	`GET /v1/models` may be used for discovery or troubleshooting
Auth header	`Authorization: Bearer YOUR_API_KEY`

Verification

Open Open Code.
Select the HaiToken provider and one of your configured models.
Send a simple prompt such as hello.
If a normal response comes back, the integration is working.

FAQ

Question	Fix
I used `/connect` but never saw a Base URL field	That is expected. `/connect` mainly stores credentials. The provider definition, Base URL, and model list belong in `opencode.json`.
`401` or authentication failed	Make sure the key is valid and that `HAITOKEN_API_KEY` is available in the same environment where Open Code is running.
`model not found`	Make sure the model name in `opencode.json` exactly matches a model ID that HaiToken exposes.
The provider was added but still does not show up	Restart Open Code after saving `opencode.json`, then check the provider picker again.

Quick start

API documentation

AI tool integration

Other

Protocol Note

What you need

Setup Steps

Step 1: Save the HaiToken API key with `/connect`

Step 2: Edit `opencode.json`

Step 3: Make sure the environment variable exists

Step 4: Restart Open Code and choose the model

Request Mapping

Verification

FAQ

Next

​Protocol Note

​What you need

​Setup Steps

​Step 1: Save the HaiToken API key with /connect

​Step 2: Edit opencode.json

​Step 3: Make sure the environment variable exists

​Step 4: Restart Open Code and choose the model

​Request Mapping

​Verification

​FAQ

​Next

Protocol Note

What you need

Setup Steps

Step 1: Save the HaiToken API key with `/connect`

Step 2: Edit `opencode.json`

Step 3: Make sure the environment variable exists

Step 4: Restart Open Code and choose the model

Request Mapping

Verification

FAQ

Next