How to Integrate Qianfan AI Models in Openclaw

Table of contents: [Show]

Qianfan is Baidu's AI platform offering powerful models like ERNIE and other Chinese-optimized LLMs. This guide walks you through integrating Qianfan models into your Openclaw setup.

Quick Start

Get a Qianfan API Key

Sign in to your Baidu Cloud account and navigate to the Qianfan Console.
Create a new application or select an existing one.
Generate an API key. The format looks like bce-v3/ALTAK-...
Copy the API key to your clipboard.

Configure Openclaw

Run the onboarding wizard to authenticate with Qianfan:

openclaw onboard --auth-choice qianfan-api-key

For more authentication options, check out our guide on LiteLLM Models Openclaw.

Set Qianfan as Your Default Model

Once authenticated, configure your default model in ~/.openclaw/openclaw.json:

{
  "agents": {
    "defaults": {
      "model": { "primary": "qianfan/ernie-bot-4" }
    }
  }
}

Available models include qianfan/ernie-bot-4, qianfan/ernie-bot-turbo, and other Baidu models.

Example Usage

After setup, you can interact with Qianfan models directly:

openclaw ask "Explain quantum computing in simple terms"

Or use it in your agent workflows for Chinese language tasks, coding assistance, or general AI interactions.

Troubleshooting

Authentication errors: Ensure your API key starts with bce-v3/ and hasn't expired.
Region issues: Qianfan works best for China-based deployments; consider latency if hosting elsewhere.
Model not found: Verify the model ID matches exactly what's available in your Qianfan console.

If you're deploying Openclaw on cloud platforms, see our Deploy Flyio Openclaw guide for production-ready setups.