User Guide

This guide is for people who need to use RePORT AI Portal. It explains what the portal does, how it helps a study team, how to set it up, and how to run one study safely. It does not explain how the code is built.

Start Here

Reader	Goal	Start with
Clinical researcher	Ask questions about a locked, PHI-scrubbed study.	Overview
Data manager	Prepare raw study files and publish a safe bundle.	Quick Start
Site operator	Install, configure, and launch the portal locally.	Installation
PI	Understand the user workflow and where output evidence lives.	Overview, then Data Pipeline

What’s Included

• Overview - what the portal is, who it helps, and when to use it.

• Installation - system requirements and setup.

• Quick Start - first run from raw study files to chat.

• Configuration - the small set of settings most users touch.

• Data Pipeline - a plain-language view of what happens when you load a study.

• Glossary - user-facing terms.

• Frequently Asked Questions - common setup, privacy, and troubleshooting questions.

Contents

Getting Started

• Overview
  • What It Does
  • Who It Helps
  • Typical Workflow
  • Privacy in Plain Language
  • When to Use It
  • Next Step
• Installation
  • Requirements
  • Install
  • Start the Web UI
  • Optional Developer Setup
  • Prepare for First Run
• Quick Start
  • Before You Start
  • 1. Place Study Files
  • 2. Choose a Model Provider
  • 3. Load the Study
  • 4. Start Chat
  • Use Existing Study
  • Common Problems
  • Next Steps

Using the Portal

• Configuration
  • Minimum Settings
  • Recommended Default
  • Study Folder
  • PHI-Related Settings
  • PHI Key
  • Where to Put More Detail
  • Next Step
• Data Pipeline
  • What “Load Study” Means
  • Input Folder
  • Output Folder
  • Running the Pipeline
  • Using an Existing Study
  • PDFs
  • Audit Files
  • Troubleshooting
• Glossary
  • Developer Terms
• Frequently Asked Questions
  • What is RePORT AI Portal?
  • Who is it for?
  • What can I ask it?
  • What files do I need?
  • Do I need an API key?
  • How do I install it?
  • How do I run it?
  • Can I use an existing processed study?
  • Does the assistant read raw files?
  • What if my PDFs may contain PHI?
  • Can I skip PDFs?
  • Where do I check what happened during a run?
  • What if I find raw PHI in output?
  • Can it handle large datasets?
  • Is there a GUI?
  • Where are developer details?

Where Details Live

User pages stay brief on purpose. They should help a study team operate the portal without needing to understand how the code is built.

Other audience-specific detail stays in:

• Developer Guide - architecture, source layout, tests, operational runbooks, and contributor guidance.

• IRB/Auditor Profile - reviewer-facing PHI handling and conformance material.