Currently, rag_api primarily uses OpenAI's embedding models for document vectorization. While OpenAI provides excellent embeddings, it would be valuable to support alternative providers, particularly Jina AI's embedding models.

Benefits of supporting Jina AI embeddings:

• Task-specific optimization through Lora fine-tuning
• Provides additional options for users with different requirements

The implementation would require minor adjustments to accommodate Jina AI's slightly different API parameters. The API examples are available at https://jina.ai/api-dashboard/embedding.

Would it be possible to consider adding Jina AI as an embedding provider option? This would enhance the flexibility of rag_api and give users more choices for their specific needs.