OpenClaw

Installation

Please ensure Node >= 22 before installation.

Run the installation command

npm install -g openclaw@latest

After installation, the following commands can be run from any directory. You do not need to switch to the OpenClaw installation directory first. If your terminal recognizes openclaw, you can copy the commands in this guide directly.

openclaw models status

If Windows reports that openclaw cannot be found, replace the leading openclaw command with your local executable path:

& "your-install-path\OpenClaw\openclaw.cmd" models status

your-install-path is a placeholder. Replace it with the parent directory where OpenClaw is installed. For example, if OpenClaw is located at D:\Apps\OpenClaw, use & "D:\Apps\OpenClaw\openclaw.cmd" models status.

Terminal Configuration

Open the terminal and run the following command to start the configuration

openclaw onboard

Select Yes

Select QuickStart

Select Skip for now

Select All providers

Choose the default model as needed; you can initially select Keep current (if using “Method 2: Manual Configuration”, it can be modified in the configuration file in the next step).

Link your personal account as needed; if not required, please select Skip for now.

Configure skills by selecting Yes.

Select npm

Configure skill selection as needed.

Select No for all options.

Select Skip for now.

Select Hatch in UI.

Configure AIHubMix

Method 1: Quick Setup via Plugin (Recommended)

Plugin source code: aihubmix-auth

Install the AIHubMix auth plugin and restart the Gateway:

openclaw plugins install @akakenle/aihubmix-auth
openclaw gateway restart

Run the auth command. Enter your AIHubMix API Key when prompted. You only need to run the command below once for any one of the four providers — the plugin uses providerAuthAliases to make the same key valid for all transports:

openclaw models auth login --provider aihubmix-openai --method api-key --set-default

The plugin organizes models into the following groups:

aihubmix-openai — gpt-* / o1-* / o3-* / o4-*, https://aihubmix.com/v1 (openai-completions)
aihubmix-anthropic — claude-*, https://aihubmix.com (anthropic-messages)
aihubmix-google — gemini-*, https://aihubmix.com/gemini/v1beta (google-generative-ai)
aihubmix-other — other models, https://aihubmix.com/v1 (openai-completions)

Shorthand model ids (e.g. claude-sonnet-4-5) are automatically routed to the correct transport by the plugin’s resolveDynamicModel hook — no need to write the full provider/model.

Prefer to skip interactive login? Set export AIHUBMIX_API_KEY=sk-... before installing — the plugin will write the key as an inline apiKey in the provider config (interactive login stores it in the auth-profile instead).

Once configured, jump to the Start section to begin using it.

Method 2: Manual Configuration

Open the ~/.openclaw/openclaw.json file and add the following configuration:

{
  "models": {
    "mode": "merge",
    "providers": {
      "aihubmix-openai": {
        "baseUrl": "https://aihubmix.com/v1",
        "api": "openai-completions",
        "models": [
          {
            "id": "gpt-4o",
            "name": "gpt-4o",
            "reasoning": false,
            "input": ["text"],
            "cost": {
              "input": 0,
              "output": 0,
              "cacheRead": 0,
              "cacheWrite": 0
            },
            "contextWindow": 128000,
            "maxTokens": 16384
          }
        ]
      },
      "aihubmix-anthropic": {
        "baseUrl": "https://aihubmix.com",
        "api": "anthropic-messages",
        "models": [
          {
            "id": "claude-sonnet-4-5",
            "name": "claude-sonnet-4-5",
            "reasoning": false,
            "input": ["text"],
            "cost": {
              "input": 0,
              "output": 0,
              "cacheRead": 0,
              "cacheWrite": 0
            },
            "contextWindow": 200000,
            "maxTokens": 64000
          }
        ]
      },
      "aihubmix-google": {
        "baseUrl": "https://aihubmix.com/gemini/v1beta",
        "api": "google-generative-ai",
        "models": []
      },
      "aihubmix-other": {
        "baseUrl": "https://aihubmix.com/v1",
        "api": "openai-completions",
        "models": []
      }
    }
  },
  "agents": {
    "defaults": {
      "model": {
        "primary": "aihubmix-anthropic/claude-sonnet-4-5"
      },
      "workspace": "your_workspace_path",
      "compaction": {
        "mode": "safeguard"
      },
      "maxConcurrent": 4,
      "subagents": {
        "maxConcurrent": 8
      }
    }
  }
}

Method 3: Switch the Default Model with Dynamic Model Commands

The AIHubMix OpenClaw plugin supports dynamic model routing. As long as a model is available in your AIHubMix account, you can call it in the provider/model-id format. The actual model availability depends on the AIHubMix console and API response. Common provider rules:

aihubmix-openai: OpenAI-compatible models, such as gpt-*, o1-*, o3-*, and o4-*
aihubmix-anthropic: Claude models, such as claude-*
aihubmix-google: Gemini models, such as gemini-*
aihubmix-other: other models called through an OpenAI-compatible endpoint, such as glm-5.2

For example, set gpt-5.4 as the default OpenClaw model:

openclaw models set aihubmix-openai/gpt-5.4

If Windows cannot find openclaw, use the local path form:

& "your-install-path\OpenClaw\openclaw.cmd" models set aihubmix-openai/gpt-5.4

Check the current default model and authentication status:

openclaw models status

Run a real call for verification:

openclaw infer model run --local --json --model aihubmix-openai/gpt-5.4 --prompt "Reply only OK"

To call a model outside the OpenAI / Claude / Gemini groups, use aihubmix-other, for example:

openclaw infer model run --local --json --model aihubmix-other/glm-5.2 --prompt "Reply only OK"

List the model catalog currently loaded by OpenClaw:

openclaw infer model list --json

Start

openclaw gateway

Test

Type hello in the chat page; if you receive a reply, the configuration is successful.

Frequently Asked Questions

Version Upgrade

openclaw plugins update @akakenle/aihubmix-auth
openclaw gateway restart

Gateway Process Occupied

openclaw gateway stop
openclaw gateway

Last updated: 2026-07-03

Coding Agents

IDE Integrations

Agent / Workflow

SDKs & Frameworks

Installation

Terminal Configuration

Configure AIHubMix

Method 1: Quick Setup via Plugin (Recommended)

Method 2: Manual Configuration

Method 3: Switch the Default Model with Dynamic Model Commands

Start

Test

Frequently Asked Questions

Version Upgrade

Gateway Process Occupied

​Installation

​Terminal Configuration

​Configure AIHubMix

​Method 1: Quick Setup via Plugin (Recommended)

​Method 2: Manual Configuration

​Method 3: Switch the Default Model with Dynamic Model Commands

​Start

​Test

​Frequently Asked Questions

​Version Upgrade

​Gateway Process Occupied

Installation

Terminal Configuration

Configure AIHubMix

Method 1: Quick Setup via Plugin (Recommended)

Method 2: Manual Configuration

Method 3: Switch the Default Model with Dynamic Model Commands

Start

Test

Frequently Asked Questions

Version Upgrade

Gateway Process Occupied