OpenAI-Compatible API Key Client Setup

Many AI clients support OpenAI-compatible APIs. In most cases, setup comes down to three fields: API Key, Base URL, and model name.

The three core fields

• API Key: created in the console and used for authentication.
• Base URL: OpenAI-compatible clients usually use https://api.cheng-zi-ai.com/v1.
• Model name: for example gpt-5.3-codex, claude-opus-4-8, or gemini-3.1-pro-high.

Cline setup idea

1. Open Cline settings.
2. Choose OpenAI Compatible as the provider.
3. Paste your API Key.
4. Set Base URL to https://api.cheng-zi-ai.com/v1.
5. Paste an available model name from the console.

ChatBox setup idea

1. Choose a custom provider.
2. Select OpenAI API as the interface type.
3. Set Base URL to https://api.cheng-zi-ai.com/v1.
4. Save and send a short test message.

NextChat setup idea

NextChat usually supports custom API endpoints. Enter the OpenAI-compatible Base URL and your API Key, then select a model.

How to know it works

• The client returns text normally.
• The console shows usage logs.
• The billed model matches your selected model.
• You no longer see 401, 404, or model-not-found errors.

Common mistakes

• Mixing Claude Code settings with OpenAI-compatible client settings.
• Forgetting /v1 in the Base URL.
• Guessing model names instead of copying them from the console.
• Using an expired or inactive key.