Question 1

What is aiiop RAS?

Accepted Answer

aiiop RAS (Retrieval-as-a-Service) is a self-hosted RAG platform that lets enterprise teams ask plain-language questions across PDF, Word, email and other document types. It retrieves relevant context and generates grounded answers using a configurable LLM.

Question 2

Can RAS run entirely on-premise without sending data to the cloud?

Accepted Answer

Yes. RAS supports self-hosted embedding models via Ollama or HuggingFace over SSH, pgvector for vector storage in your existing PostgreSQL database, and local LLMs such as Llama 3 and Mistral via Ollama. No data needs to leave your infrastructure.

Question 3

How does RAS handle sensitive documents with PII?

Accepted Answer

RAS integrates Microsoft Presidio for PII detection and redaction. Redacted versions of text are used for embedding and LLM context. Original files remain intact in storage. PII scanning runs at pipeline build time, not on every query.

Question 4

What document types does RAS support?

Accepted Answer

RAS supports PDF, DOCX, TXT, CSV, JSON, XML, HTML, Markdown, EML, MSG, images (via OCR) and code files. Documents can be uploaded directly or synced from S3, Azure Blob Storage or Google Cloud Storage.