Investigate and Integrate `rag_api` (LibreChat RAG Service)

[stephane-segning]

Summary

As a platform engineer, I want to investigate the rag_api project (Danny Avila) to understand how it works, how LibreChat integrates with it, and whether we can reuse it for other applications beyond LibreChat.

Acceptance Criteria

Given the rag_api repository and LibreChat RAG documentation,
when we complete the investigation,
then we can decide whether to adopt it as a standalone RAG microservice in our platform.

• It’s done when we understand the deployment model, API surface, and storage backends of rag_api.
• It’s done when we can run it locally and connect LibreChat successfully.
• It’s done when we document whether it can serve other clients besides LibreChat.

Additional Sections

• Design: N/A (investigation + architecture notes).

• Technical Information:

  • Repo: https://github.com/danny-avila/rag_api
  • LibreChat RAG API docs: https://www.librechat.ai/docs/configuration/rag_api
  • LibreChat feature docs: https://www.librechat.ai/docs/features/rag_api

• Notes: Evaluate multi-tenancy, auth integration, and vector DB support.

• ToDo: @clarify whether we host it or replace it with our own RAG service.

Definition of Done

• A written integration report exists.
• Proof-of-concept connection to LibreChat works.
• Decision recorded: adopt / fork / replace.