☰ Table of Contents

Welcome to the RAGU Help Center

Your guide to using the secure and offline Semantic Assistant.

Quick Start

Get RAGU up and running in three simple steps:

1. Acquire and Validate: Enter your License Key (requires internet access only once).
2. Download a Model: Navigate to Settings → Manage Models and download your preferred LLM model (requires internet).
3. Start Chatting: Use the Library tab to ingest your first file and start a private, offline conversation!

How to Acquire a License

Detailed instructions on how to purchase and activate your license key. Remember, your key ensures exclusive and secure access to your data.

Access the website RAGU-PRO.COM and purchase your license. After payment confirmation, an email will be sent to you containing the key, which must be entered as shown in the image below.

The Differences Between RAGU-Pro and Freemium

Library, Categories, and Documents

The Library is where all your private files are managed. You can Create, Read, Update, and Delete (CRUD) documents. Organization is handled through user-defined Categories.

All files within a category must be in the same language.

• File Ingestion: How to upload PDFs, text files, or short books for RAG.

Chat / RAG (Retrieval-Augmented Generation)

This is the core feature. Select a previously created category to start a private conversation session about the document.

Encryption: All chat sessions are encrypted and stored only locally on your machine.

Summary and Translation

Use this tool for quick text operations:

• Summarize: Summarize long texts or short books efficiently using the selected LLM.
• Translation: Fast, offline translation between languages supported by your RAGU.

Dictation, Synthetic Reading, and Multi-format Formatting

Leveraging specialized offline tools for voice and text manipulation.

1. Dictation: Used for fast, local, and private speech recognition.
2. Synthetic Reading: Utilizes the System's TTS (Text-to-Speech) feature to read documents aloud.
3. Formatting: Options to clean, style, and adjust input text before summarization or chat.
4. Export: Options for various formats or to the Clipboard.

Settings Overview

General application configuration, including language settings (American English / Brazilian Portuguese), password changes, and user maintenance (Administrator Only).

Settings → Manage Models

The control center for your AI engine.

• RAGU Pipeline: RAGU uses a pipeline system to manage and serve models.
• Download Models: Instructions on how to select and download models through the integrated interface (Internet required).
  With RAGU, you have the flexibility to swap models based on your task (e.g., speed vs. complexity).
• CPU ONLY: RAGU was designed to use CPU only, eliminating the need for expensive GPUs. Performance will depend on the local CPU.

In the right pane of the tab, you can search for a new model to download. Look for models that feature "completion", "thinking", and/or "chat".

The quality of the models depends on the number of parameters, 1.5b, 7b...671b; the more parameters (and file size), the greater the "intelligence" of the model, and also the higher the consumption of your computer's resources.

The more memory (RAM) and processing power, the faster the model's responses.

Copy the name of the chosen model and paste it into the text box indicated in the image above.