Aquiles-ai - Aquiles-RAG: A high-performance RAG server
by•
Aquiles integrates AI across enterprises and delivers high‑performance Retrieval‑Augmented Generation (RAG) on Redis. Multi‑provider AI, enterprise‑ready, and easy integration.
Replies
Best
Maker
📌
I’ve been developing Aquiles-RAG for about a month. It’s a high-performance RAG server that uses Redis as the vector database and FastAPI for the API layer. The project’s goal is to provide a production-ready infrastructure you can quickly plug into your company or AI pipeline, while remaining agnostic to embedding models — you choose the embedding model and how Aquiles-RAG integrates into your workflow.
What it offers
- An abstraction layer for RAG designed to simplify integration into existing pipelines.
- A production-grade environment (with an Open-Source version to reduce costs).
- API compatibility between the Python implementation (FastAPI + Redis) and a JavaScript version (Fastify + Redis — not production ready yet), sharing payloads to maximize compatibility and ease adoption.
Why I built it
I believe every RAG tool should provide an abstraction and availability layer that makes implementation easy for teams and companies, letting any team obtain a production environment quickly without heavy complexity or large expenses.
Love that you made Aquiles-RAG embedding-model agnostic—plug and play with whatever works for your team is super smart, tbh. Open-source too? That’s wild!
Report
Maker
@cruise_chen THANK YOU VERY MUCH. The idea behind the project was to be completely agnostic to the embedding model from the start, so it can be integrated much more easily into your pipelines without much trouble. I hope you like it and that it adds a lot of value!
This is truely awesome! The whole "choose your own embedding model" thing is kinda genius imo — so many RAG tools lock you in, right? Being able to just plug it into my existing workflow is a huge win. How customizable is it for different types of embeddings?
Report
Maker
A BIG THANK YOU TO EVERYONE ON THE Aquiles-RAG LAUNCH, we've gone from 3 stars to 14 stars on GitHub, thank you so much for the feedback and support :D
Report
Maker
Hey hello everyone, quick update we've released a short video on how to deploy Aquiles-RAG in Render :D Video:
Report
Maker
We are improving and implementing new things to improve RAG recovery and maintain backward compatibility🥳💥
Report
Maker
HEYYYY there is a new version of Aquiles-RAG that includes internal speed improvements and adds new security layers to obtain a better result in both indexing and searching 🥳🔨💥⛏️
Replies
Agnes AI
Love that you made Aquiles-RAG embedding-model agnostic—plug and play with whatever works for your team is super smart, tbh. Open-source too? That’s wild!
@cruise_chen THANK YOU VERY MUCH. The idea behind the project was to be completely agnostic to the embedding model from the start, so it can be integrated much more easily into your pipelines without much trouble. I hope you like it and that it adds a lot of value!
GPT-4o
This is truely awesome! The whole "choose your own embedding model" thing is kinda genius imo — so many RAG tools lock you in, right? Being able to just plug it into my existing workflow is a huge win. How customizable is it for different types of embeddings?
A BIG THANK YOU TO EVERYONE ON THE Aquiles-RAG LAUNCH, we've gone from 3 stars to 14 stars on GitHub, thank you so much for the feedback and support :D
Hey hello everyone, quick update we've released a short video on how to deploy Aquiles-RAG in Render :D Video:
We are improving and implementing new things to improve RAG recovery and maintain backward compatibility🥳💥
HEYYYY there is a new version of Aquiles-RAG that includes internal speed improvements and adds new security layers to obtain a better result in both indexing and searching 🥳🔨💥⛏️