|
<< Click to Display Table of Contents >> Raynet One > 2026.1 > User Guide > First impressions > Recurring areas and elements > AI Chat Technical Architecture |
This section describes the internal technical architecture of the AI Chat feature, including its core components, communication flow, external service integrations, and security principles.
The AI Chat uses a centralized server-side storage model.
Key characteristics:
•All chat sessions are stored in the backend database.
•Persistent context handling across sessions.
•Support for multiply parallel chat sessions per user.
•Audit logging and traceability.
•Token usage tracking.
•Multiple chat session.
•No local browser storage dependency.
This ensures reliability, traceability, and full server control over chat data.
The AI Chat uses WebSocket technology to provide real-time responses. When you send a message:
1.Your message is sent to the AI service.
2.The assistant processes your request and may query the database if needed.
3.You will see status updates showing what the assistant is working on.
4.The response streams back to you in real-time.
The frontend does not communicate directly with any AI provider.
Instead, the communication flow is:
Frontend < Backend < LiteLLM < AI Provider
Key principles:
•Frontend never talks directly to AI provider.
•Backend handles all AI requests.
•All AI requests are handled server-side.
•API keys and credentials remain on the server.
•Request validation & sanitization before provider call.
•Rate limiting & retry logic.
•Centralized Logging & cost tracking.
•Abstraction layer (LiteLLM).
•Full control over request orachestration.
This architecture ensures security, maintainability, and provider independence.
LiteLLM acts as an abstraction layer between Raynet One and external AI providers.
Responsibilities:
•Provider abstraction (OpenAI, Azure, etc.)
•AI Model switching without application code changes
•Centralized model configuration
•secure API key management
•Usage & cost metrics collection
This design allows flexible provider switching and centralized AI governance.
Voice capabilities are integrated via external speech services.
Technical components:
•Speech-to-Text (STT) via Azure Speech API
•Text-to-Speach (TTS) via Azure Speech API
•Server-side processing and orchestration
•Browser independent voice handling
•Custom TTS provider instead of browser-native speechSynthesis
•Voice command handling logic (stop, pause, resume)
The voice layer is fully integrated into the backend-controlled AI communication flow.
Voice capabilities are optional and remain unavailable unless the Speech integration is configured. Configuration is performed via the Voice Control integration settings (see AI Chat Configuration & Deployment > Voice Mode Configuration). The voice layer follows the same backend-controlled communication flow, security principles, and permissions boundaries as text-based interaction.
The License Management workspace relies on multiple external services to provide AI-supported functionality and enriched data processing capabilities.
All services listed below are mandatory components. The License Management workspace is not operational unless these integrations are fully configured.
SAMCloud Integration
SAMCloud provides AI-enhanced services specifically for License Management.
Responsibilites:
•Document parsing (e.g., invoices, license contracts)
•Extraction of structured core information from uploaded files
•Access to the Encyclopedia service
•Additional AI-driven analysis services
Configuration (Docker example - sanitized):
SAMCloud__Url: "<SAMCLOUD_URL>"
SAMCloud__ApiKey: "<SAMCLOUD_API_KEY>"
SAMCloud__MaxAttempts: 5
MaxPollingAttempts: 6
PollingDelaySeconds: 10
Without SAMCloud, License management remains operational but lacks automated document analysis and enriched AI capabilities.
Technology Catalog Integration
The Technology Catalog provides structured product and technology data used within License Management and AI-assisted workflows.
Responsibilities:
•Enriched product metadata
•Technology normalization
•Structured software and vendor information
•Data foundation for AI-supported analysis
Configuration (Docker example - sanitized):
Catalog__Url: "<CATALOG_SERVICE_URL>"
Catalog__ApiKey: "<CATALOG_API_KEY>"
Catalog__MaxAttempts: 3
The Catalog service enhances the quality and consistency of license and product-related insights.
SAMCloud provides AI-enhanced services specifically for License Management.
AI Chat /LLM Integration
AI Chat enables natural language interaction and AI-supported analysis within the system.
Responsibilites:
•Natural language query handling
•Context-aware responses
•AI-assisted evaluation of license and asset data
The AI Chat backend communicates with the configured AI provider through the LiteLLM abstraction layer.
Configuration (Docker example - sanitized):
AiConfig__Url: "<AI_SERVICE_URL>"
AiConfig__ApiKey: "<AI_API_KEY>"
AiConfig__Model: "<MODEL_NAME>"
Architectural Summary
For full License Management capabilities:
•SAMCloud enables document intelligence and AI-enhanced parsing
•The Technology Catalog provides structured product intelligence
•AI Chat enables conversational interaction and AI-assisted analysis
Together, these services from the AI supported License Management ecosystem.