OpenClaw API-KEY Configuration

1. Course Overview

This course will detail how to configure the AI model API-KEY for OpenClaw, including the application process for API-KEYs, two configuration methods (CLI command configuration and configuration file modification), and how to verify if the configuration is successful. After completing this course, you will be able to successfully connect to and use various AI large model service providers.

1.1 Course Content

• Understand the types of model service providers supported by OpenClaw
• Master the method for applying for API-KEYs
• Learn how to view and verify the status of connected models

2. API-KEY Application

2.1 Recommended Model Service Providers

• Recommended for beginners: Refer to the accompanying tutorial [Chapter 8 Online AI Large Model Development] — [Online Large Models (Text Interaction)] — [2. Tongyi Qianwen Large Model] to apply for the corresponding model provider API-KEY.

• OpenClaw supports numerous model service providers, including:

  • OpenAI - GPT series (API + Codex)
  • Anthropic - Claude series (API + Claude Code)
  • Google - Gemini series
  • Alibaba Cloud - Qwen Cloud, Tongyi Qianwen
  • DeepSeek - DeepSeek models
  • Zhipu AI - GLM models
  • Moonshot AI - Kimi + Kimi Coding
  • Ollama - Local models (cloud + local)
  • And many other service providers

• You can freely choose a model service provider based on your actual needs and regional network policies, register on the corresponding official website, and apply for an API-KEY.

3. Configuring OpenClaw API-KEY

3.1 Configuration Methods

OpenClaw provides two ways to configure API-KEYs:

1. CLI Command Configuration (suitable for beginners)

2. Directly Modify Configuration File (suitable for developers)

   • Configuration file path: $HOME/.openclaw/openclaw.json

3.2 Quick Configuration

• Suitable for users who want to get started quickly. You need to register for an Alibaba Cloud Bailian large model account and create an API-KEY first (refer to the accompanying tutorial [Chapter 8 Online AI Large Model Development] — [Online Large Models (Text Interaction)] — [2. Tongyi Qianwen Large Model]).
• The factory-preset configuration file has already configured the Bailian model parameters; Simply fill in your API-KEY and baseUrl to start using it directly.
• Open the OpenClaw configuration file:

xxxxxxxxxx
nano $HOME/.openclaw/openclaw.json

xxxxxxxxxx
nano $HOME/.openclaw/agents/main/agent/models.json

3.2.1 Configure baseUrl

• Users need to configure the tongyi_base_url corresponding to their account. Go to the Model Service Platform Bailian Console to apply for an API key. Copy the API key and OpenAI-compatible address (Base URL), then fill them into the corresponding fields.

3.2.2 Fill in apiKey and baseUrl

• After filling it in, press Ctrl+X, then press Y to confirm saving, and press Enter to exit.

4. View Connected Model Status

4.1 View the List of Configured Models

After configuration, you can use the following command to view the connected models:

xxxxxxxxxx
openclaw models list

4.2 Verify Configuration Success

1. Test Conversation

   xxxxxxxxxx
   openclaw tui
   

   Start the TUI interface, send a test message (e.g., "hello"), and confirm that the model responds normally.

5. Other Access Methods (Advanced)

• If you have your own model provider or need to access other model providers later, refer to the following tutorial.

5.1 CLI Command Configuration for API-KEY (Suitable for Beginners)

1. Start Configuration Wizard

   xxxxxxxxxx
   openclaw onboard
   

• Select [Yes] and press Enter.

• Select QuickStart for the configuration mode and press Enter.

2. Select Model Service Provider

   • Use the up and down arrows to select the service provider for which you have applied for an API-KEY, and press Enter to confirm. Here we take Alibaba Cloud Bailian model provider as an example; the configuration method for other model providers is the same.
   • Here select Custom Provider [Custom Provider]

   

   • Enter the base_url of the model provider.

3. Enter API-KEY

• After selecting the provider, the system will prompt you to enter the corresponding API-KEY.

• Paste or type the API-KEY you applied for, and press Enter to confirm.

• Select OpenAI-compatible for the interface protocol.

4. Fill in the model name, taking MiniMax as an example.

• Endpoint ID is the model provider identifier and can be filled in freely.

• Model alias (optional) can be left blank.

5. Other Configurations

• For other configuration items, select Skip for all as shown below; no configuration is needed.

• For Enable Hooks, press space to select Skip for now.

• Select Restart to restart the gateway and apply the configuration.

• Close the popped-up OpenClaw webpage.

5.2 Modify Configuration File (Recommended Method)

Configuration File Location

Configuration files are located at: $HOME/.openclaw/openclaw.json

xxxxxxxxxx
                       `$HOME/.openclaw/agents/main/agent/models.json`

Edit Configuration File

1. Open Configuration File 1

   xxxxxxxxxx
   nano $HOME/.openclaw/openclaw.json
   # Or use your preferred editor, such as vim, code, etc.
   

2. Add Model Provider Configuration

   Add or modify the models.providers section in the configuration file:

   xxxxxxxxxx
   {
     "models": {
       "providers": {
         "openai": {
           "apiKey": "OPENAI_API_KEY"
         },
         "anthropic": {
           "apiKey": "ANTHROPIC_API_KEY"
         },
         "google": {
           "apiKey": "GOOGLE_API_KEY"
         }
       }
     },
     "agents": {
       "defaults": {
         "model": {
           "primary": "openai/gpt-4"
         }
       }
     }
   }
   

3. Open Configuration File 2

   xxxxxxxxxx
   nano $HOME/.openclaw/agents/main/agent/models.json
   

4. Add Model Provider Configuration

   Add or modify as follows:

   

5. Configuration Description

   • models.providers: Configure the API-KEYs for each model service provider.

     • openai.apiKey: OpenAI's API-KEY
     • anthropic.apiKey: Anthropic's API-KEY
     • google.apiKey: Google's API-KEY

   • agents.defaults.model.primary: Set the default model to use.

     • Format is provider/model_name, e.g., openai/gpt-4, anthropic/claude-3, google/gemini-pro

6. Save and Exit

   • Be careful not to randomly delete periods or add other symbols when modifying.
   • Press Ctrl+O to save, Ctrl+X to exit (nano editor).

7. Factory Default Bailian Configuration (Can be directly modified or referenced)

xxxxxxxxxx
"models": {
    "mode": "merge",
    "providers": {
    "bailian": {
      "baseUrl": "https://dashscope.aliyuncs.com/compatible-mode/v1",
      "apiKey": "sk-a98b519b6bab4e35a7b82b4a95b8a4f6",
      "api": "openai-completions",
      "models": [
        {
          "id": "qwen3.5-plus",
          "name": "qwen3.5-plus",
          "reasoning": false,
          "input": [
            "text",
            "image"
          ],
          "contextWindow": 1000000,
          "maxTokens": 65536,
          "cost": {
            "input": 0,
            "output": 0,
            "cacheRead": 0,
            "cacheWrite": 0
          }
        },
        {
          "id": "MiniMax-M2.5",
          "name": "MiniMax-M2.5",
          "reasoning": false,
          "input": [
            "text"
          ],
          "cost": {
            "input": 0,
            "output": 0,
            "cacheRead": 0,
            "cacheWrite": 0
          },
          "contextWindow": 196608,
          "maxTokens": 32768
        },
        {
          "id": "glm-4.7",
          "name": "glm-4.7",
          "reasoning": false,
          "input": [
            "text"
          ],
          "cost": {
            "input": 0,
            "output": 0,
            "cacheRead": 0,
            "cacheWrite": 0
          },
          "contextWindow": 202752,
          "maxTokens": 16384,
          "compat": {
            "thinkingFormat": "qwen"
          }
          }
        ]
      }

6. Comm