RAGapp local setup with Docker and OpenAI

This tutorial, uploaded on May 25, 2024, walks through a local RAGapp installation and a document-based chat workflow. The presenter uses Ubuntu 22.04 and Docker 26.1.3. No RAGapp release number is supplied; the demonstration uses a Docker image tagged latest, so the setup reflects the upload period.

After cloning the repository and starting the container, the presenter opens localhost on port 8000 and enters an OpenAI API key. This is a self-hosted application using a hosted model in the demonstrated configuration. The presenter also names Gemini and Ollama as provider options, but reports problems with the Ollama integration and postpones that walkthrough. His suggestion of at least 16 GB of VRAM for Ollama is advice, not a tested minimum in this video.

The workflow includes editing the assistant prompt, uploading Oracle database documentation, and asking questions about it. The interface displays retrieval events and links to the uploaded file as an answer source. A separate chat page opens through the Start the app control, providing a URL for users.

The presenter discusses Kubernetes deployment on company servers or cloud infrastructure. He says the application lacked authentication and authorization at the time and would require an additional access-control layer. The video does not demonstrate that deployment.