Skip to main content

​
What is OpenClaw

OpenClaw is an open-source local AI gateway tool that runs in the terminal, supporting unified access to multiple AI model providers. It has built-in support for mainstream providers like OpenAI, Anthropic, and Google, and also supports connecting any OpenAI or Anthropic compatible third-party endpoint via Custom Provider. By configuring LaoZhang API as a Custom Provider, you can access 200+ AI models in OpenClaw, including the latest GPT-5, Claude Opus 4.6, Gemini 2.5 Pro, and more.

​
Prerequisites

​
Configuration Steps

​
Step 1: Select Custom Provider

After launching the OpenClaw configuration wizard, scroll to the bottom of the Model/auth provider list and select Custom Provider (Any OpenAI or Anthropic compatible endpoint). Select Custom Provider as model provider in OpenClaw
Custom Provider supports any endpoint compatible with OpenAI or Anthropic interfaces. LaoZhang API is fully compatible with the OpenAI API format.

​
Step 2: Enter API Base URL

In the API Base URL field, enter the LaoZhang API endpoint: 
http://api.laozhang.ai/v1
Enter LaoZhang API Base URL in OpenClaw
You can also use https://api.laozhang.ai/v1. HTTPS is recommended for better security.

​
Step 3: Choose API Key Method

When asked how to provide the API key, select Paste API key now to store the key directly in OpenClaw config. Choose to paste API key directly in OpenClaw

​
Step 4: Enter API Key

Paste your API key from the LaoZhang API console in the API Key field (a string starting with sk-). Enter LaoZhang API key in OpenClaw
API keys can be created at LaoZhang API Token Management. When creating a new token, select pay-per-use billing, default group, and keep all other settings as default.

​
Step 5: Select Endpoint Compatibility

Under Endpoint compatibility, select OpenAI-compatible (Uses /chat/completions). Select OpenAI-compatible endpoint mode in OpenClaw LaoZhang API is fully compatible with OpenAI’s /chat/completions endpoint. Select this option to proceed.

​
Step 6: Set Model ID

Enter the model name you want to use in the Model ID field. Set AI model ID in OpenClaw

​
Step 7: Verify and Complete

OpenClaw will automatically verify the configuration. Once verification succeeds, the system generates an Endpoint ID (e.g., custom-api-laozhang-ai) and the setup is complete. OpenClaw verification successful with generated Endpoint ID
Verification may return an error on the first attempt. Simply retry and it should succeed. If verification keeps failing, select Change model and try a different Model ID.
Through LaoZhang API, you can use the following popular models in OpenClaw:
Model SeriesModel IDHighlights
GPT-5gpt-5.2OpenAI’s latest flagship model
Claude Opus 4.6claude-opus-4-6Anthropic’s most capable reasoning model
Claude Sonnet 4.6claude-sonnet-4-6Balanced performance and cost
Gemini 2.5 Progemini-2.5-proGoogle’s latest multimodal model
GPT-4ogpt-4oClassic, stable, and fast
Model Selection Tips
  • Complex coding tasks: Use claude-opus-4-6 or gpt-5.2
  • Daily conversation: Use gpt-4o
  • Multimodal tasks: Use gemini-2.5-pro
  • Best value: Use claude-sonnet-4-6

​
Configuration Summary

After completing setup, your OpenClaw key parameters are:
SettingValue
ProviderCustom Provider
API Base URLhttp://api.laozhang.ai/v1
Endpoint CompatibilityOpenAI-compatible
Endpoint IDcustom-api-laozhang-ai

​
Troubleshooting

​
Verification Returns Error

Verification may return an error on the first attempt. Simply retry and it usually succeeds.

​
Connection Timeout

  • Check your network connection
  • Verify the API Base URL is entered correctly
  • Try using https://api.laozhang.ai/v1

​
Model Call Failures

  • Confirm your account has sufficient balance
  • Check that the Model ID is spelled correctly
  • Review call logs on the LaoZhang API Log page for troubleshooting
Need more help? Visit LaoZhang API Official Website.