[Question] Why so many specific backend dependencies ?

[Greatz08]

Instead of just supporting OpenAI you can try to integrate this project with litellm or similar open source project which can give people easy access to 100+ LLMs easily, be it open source or closed source.

(LiteLLM is an open‐source gateway that lets you call over 100 LLM APIs (like OpenAI, Anthropic, Azure, and more) using a consistent, OpenAI-like format. It abstracts away provider-specific complexities by standardizing inputs, outputs, error handling, and logging, so you can switch models without rewriting your code. Additionally, it offers built-in features like streaming support, load balancing, and spend tracking, which help create scalable and cost-efficient AI applications.)

Why pinecone only when there are different open source vector db's that people can selfhost for there personal usage and there are other open source selfhostable options available which can be supported. Taking this proprietary approach to make things easier for many to run is good but again you can support or add this open source selfhostable project info so that those who dont want to be dependent on all closed source projects can explore open source selfhostable path.