- Add support for Gemini, OpenAI, and Ollama providers
- Set default models (gemini-flash-1.5 for Gemini, gpt4o-mini for OpenAI)
- Implement provider factory pattern
- Update schema handling for each provider
- Add environment variable configuration
- Maintain token tracking across providers
Co-Authored-By: Han Xiao <han.xiao@jina.ai>
