AI Model Configuration Guide
SuperCoder needs to be configured for the AI model to work properly, and offers two main ways to configure it:
Approach 1: Using the OpenAI API (requires internet connection)
- Register on the OpenAI website and get your API key.
- Setting environment variables:
export OPENAI_API_KEY=你的密钥 - Optionally specify the model:
export OPENAI_MODEL=模型名称
Approach 2: Use native or compatible APIs (can be used offline)
- Set the base URL:
export SUPERCODER_BASE_URL=模型地址 - Configure the API key:
export SUPERCODER_API_KEY=密钥 - Specifies the model name:
export SUPERCODER_MODEL=模型名 - Additional settings are required for Google Gemini:
export SUPERCODER_GEMINI_MODE=true
Important Notes
- Local models need to support tool call functionality
- It is recommended to test small tasks first to validate the model response
- API calls may incur fees, so be careful to monitor usage.
- Model performance will directly impact tool performance
This answer comes from the articleSuperCoder: Intelligent Code Assistant for Command Line Terminal RunningThe
May not be reproduced without permission:AI productivity tools " How do I configure the AI models used by SuperCoder?
