Using TOKENOPEN API with OpenClaw

OpenClaw is an open-source AI coding agent. It can autonomously read files, write code, execute commands, and complete complex development tasks. By configuring TOKENOPEN API, you can use all models available on the platform.

OpenClaw requires Node.js 22+. Please ensure Node.js is installed before proceeding.

Step 1: Install OpenClaw

Use the official install script:

curl -fsSL https://install.openclaw.dev | bash

Verify the installation:

openclaw --version

Run in PowerShell:

iex (irm https://install.openclaw.dev/windows)

Verify the installation:

openclaw --version

Step 2: Configure TOKENOPEN API

Get an API Key

(Optional) Install System Daemon

If you want OpenClaw to run persistently as a background service, run:

openclaw onboard --install-daemon

Edit the Configuration File

OpenClaw's configuration file is located at ~/.openclaw/openclaw.json. Open and edit it:

mkdir -p ~/.openclaw
vim ~/.openclaw/openclaw.json

Paste the following content:

{
  "models": {
    "mode": "merge",
    "providers": {
      "my-proxy": {
        "baseUrl": "replace-based-on-the-table-below",
        "apiKey": "replace-with-your-api-key",
        "api": "replace-based-on-the-table-below",
        "authHeader": true,
        "models": [
          {
            "id": "replace-with-model-id",
            "name": "replace-with-model-name",
            "contextWindow": 128000,
            "maxTokens": 8192
          }
        ]
      }
    }
  },
  "agents": {
    "defaults": {
      "model": {
        "primary": "replace-based-on-the-table-below"
      }
    }
  }
}

Reference options:

Parameter	Example value	Notes
`models[].id`	`gpt-5.5` or `claude-sonnet-4-6`	Set the actual model ID to call, matching a model identifier supported by TOKENOPEN
`baseUrl`	`https://www.tokenopen.ai/v1` or `https://www.tokenopen.ai`	The former is for `openai-completions`; the latter is for `anthropic` and should not include `/v1`
`api`	`openai-completions` or `anthropic`	Choose the value that matches the API type you use
`agents.defaults.model.primary`	`gpt-5.5` or `claude-sonnet-4-6`	Set this to the primary model ID you want to use

Press Esc, then type :wq! to save and quit.

mkdir -p ~/.openclaw
vim ~/.openclaw/openclaw.json

Paste the following content:

{
  "models": {
    "mode": "merge",
    "providers": {
      "my-proxy": {
        "baseUrl": "replace-based-on-the-table-below",
        "apiKey": "replace-with-your-api-key",
        "api": "replace-based-on-the-table-below",
        "authHeader": true,
        "models": [
          {
            "id": "replace-with-model-id",
            "name": "replace-with-model-name",
            "contextWindow": 128000,
            "maxTokens": 8192
          }
        ]
      }
    }
  },
  "agents": {
    "defaults": {
      "model": {
        "primary": "replace-based-on-the-table-below"
      }
    }
  }
}

Reference options:

Parameter	Example value	Notes
`models[].id`	`gpt-5.5` or `claude-sonnet-4-6`	Set the actual model ID to call, matching a model identifier supported by TOKENOPEN
`baseUrl`	`https://www.tokenopen.ai/v1` or `https://www.tokenopen.ai`	The former is for `openai-completions`; the latter is for `anthropic` and should not include `/v1`
`api`	`openai-completions` or `anthropic`	Choose the value that matches the API type you use
`agents.defaults.model.primary`	`gpt-5.5` or `claude-sonnet-4-6`	Set this to the primary model ID you want to use

Press Esc, then type :wq! to save and quit. Or open in TextEdit:

open -e ~/.openclaw/openclaw.json

Run in PowerShell to create the config file automatically:

New-Item -ItemType Directory -Force -Path "$env:USERPROFILE\.openclaw" | Out-Null

@'
{
  "models": {
    "mode": "merge",
    "providers": {
      "my-proxy": {
        "baseUrl": "replace-based-on-the-table-below",
        "apiKey": "replace-with-your-api-key",
        "api": "replace-based-on-the-table-below",
        "authHeader": true,
        "models": [
          {
            "id": "replace-with-model-id",
            "name": "replace-with-model-name",
            "contextWindow": 128000,
            "maxTokens": 8192
          }
        ]
      }
    }
  },
  "agents": {
    "defaults": {
      "model": {
        "primary": "replace-based-on-the-table-below"
      }
    }
  }
}
'@ | Set-Content -Path "$env:USERPROFILE\.openclaw\openclaw.json" -Encoding UTF8

Reference options:

Parameter	Example value	Notes
`models[].id`	`gpt-5.5` or `claude-sonnet-4-6`	Set the actual model ID to call, matching a model identifier supported by TOKENOPEN
`baseUrl`	`https://www.tokenopen.ai/v1` or `https://www.tokenopen.ai`	The former is for `openai-completions`; the latter is for `anthropic` and should not include `/v1`
`api`	`openai-completions` or `anthropic`	Choose the value that matches the API type you use
`agents.defaults.model.primary`	`gpt-5.5` or `claude-sonnet-4-6`	Set this to the primary model ID you want to use

Or open the config file manually in Notepad:

notepad "$env:USERPROFILE\.openclaw\openclaw.json"

Set apiKey to the key obtained from the TOKENOPEN Console. You can adjust baseUrl, api, and agents.defaults.model.primary based on the table above. You can also add multiple model objects in the models array if needed.

Step 3: Use OpenClaw

Start a Chat

Once configured, start a conversation directly from the command line:

openclaw chat

Common Commands

Command	Description
`openclaw chat`	Start a conversation
`openclaw chat --session <session-id>`	Start a new session
`openclaw dashboard`	Open the web management UI
`openclaw models default set [model]`	Set the default model
`openclaw --help`	View all available commands

Supported models on TOKENOPEN (examples):

Model ID	Description
`gpt-5.5`	GPT-5.5, latest flagship model
`gpt-5.4-mini`	GPT-5.4 Mini, low-latency lightweight model
`claude-sonnet-4-6`	Claude Sonnet 4.6, balanced model
`deepseek-r1`	DeepSeek R1, strong reasoning model

For the full model list and pricing, visit the Model Marketplace.

Using TOKENOPEN API with OpenClaw

On this page