AYITA User Guide

AYITA is designed to provide secure, intelligent, and hyper-personalized AI assistance across multiple domains. This guide covers key user functionalities, including team collaboration, secure dialogues, fine-tuning, knowledge management, and personalization.

1. Getting Started with AYITA

### Accessing AYITA - AYITA is available as a web-based and self-hosted application. - Login is required via secure authentication (Keycloak-based). - Users can set up their personal workspace, choose a default language, and configure privacy settings.

### AI Interaction Basics

Users can interact with AYITA through a chat interface, with full support for:

  • Natural language understanding (context-aware responses).

  • Task and knowledge-based queries.

  • Real-time processing of structured and unstructured data.

Responses are context-aware, meaning AYITA adapts to user preferences and remembers key interactions.

2. Secure Dialogues & Privacy Features

### End-to-End Encrypted Conversations - All user interactions in AYITA are end-to-end encrypted, ensuring data confidentiality. - Chat history can be disabled or stored locally (based on user preferences). - Users can set self-destructing messages, limiting conversation retention.

### Private Mode & Local AI Processing - AYITA supports fully private interactions, running on local LLM instances. - No data is sent to external servers, ensuring complete control over user conversations. - Memory can be toggled off for temporary discussions without retention.

3. Using LoreBook for Knowledge Management

### What is LoreBook? - LoreBook is AYITA’s internal memory module, storing key facts, user preferences, and structured knowledge. - Users can add facts manually or allow automatic fact extraction.

### Adding & Managing Facts

Users can create custom fact entries for personalization and context retention.

Example usage:

  • Remembering company-specific policies.

  • Storing client preferences and previous requests.

  • Maintaining a list of project details for future reference.

Facts are retrievable on-demand, meaning AYITA will refer to stored facts when responding to queries.

4. Fine-Tuning for Personalized AI

### Using Fine-Tuning in AYITA

AYITA allows users to adapt AI behavior through Parameter-Efficient Fine-Tuning (PEFT) and Half Fine-Tuning.

Fine-tuning is useful for:

  • Adapting AI responses to specific corporate terminology.

  • Training models on company-specific documentation.

  • Improving accuracy for domain-specific inquiries.

### Training AI on Custom Data - Users can upload datasets and train models locally. - AYITA supports on-the-fly fine-tuning, meaning models adapt without full retraining.

5. Retrieval-Augmented Generation (RAG)

### What is RAG in AYITA? - RAG enhances AI-generated responses by retrieving relevant data in real time. - Unlike traditional AI models, RAG ensures responses are grounded in verified sources.

### Using RAG for Document Search

AYITA allows users to upload corporate documents and query specific information.

Example use cases:

  • Searching legal precedents in a law firm’s knowledge base.

  • Extracting insights from financial reports.

  • Finding relevant sections in a company’s internal documentation.

6. Personalizing AYITA with Avatar & Settings

### Configuring the Virtual Avatar

AYITA includes a customizable avatar, allowing users to adjust voice, appearance, and personality traits.

The avatar can:

  • Reflect a professional assistant style (corporate mode).

  • Act as a friendly, interactive guide (casual mode).

### User Settings & Customization

Users can configure:

  • Memory settings (enable/disable long-term context storage).

  • AI behavior (formal/informal tone).

  • Integration settings (connecting AYITA to external tools).

7. Workgroup & Enterprise Use Cases

### Collaboration in Teams

AYITA supports multi-user collaboration, making it ideal for:

  • Cross-departmental task coordination.

  • Internal Q&A systems for corporate employees.

  • Automated knowledge retrieval in organizations.

Teams can share customized AI models, ensuring uniform knowledge and assistance across departments.

### AYITA for Professionals & Researchers

Professionals use AYITA for:

  • Legal document summarization.

  • Medical knowledge retrieval.

  • Financial analysis and risk assessment.

Researchers benefit from AI-powered knowledge synthesis, helping them process academic papers and reports.

8. Getting Support & Future Updates

  • For troubleshooting and FAQs, visit the [AYITA Documentation](system_guide.html).

  • For customization and integrations, check the [Developer Guide](developer_guide.html).

  • For support or feedback, visit the [AYITA Community](contact.html).

—

Next Steps:

Learn how to deploy AYITA in your environment: See the [System Guide](system_guide.html).