Skip to content

Release Notes: uxopian-ai v2025.0.0

Release Date: December 2025 Version: 2025.0.0

We are thrilled to announce the general availability of uxopian-ai 2025.0.0!

uxopian-ai is a complete, standalone framework designed to accelerate the integration of GenAI features into enterprise applications. Built on Java 21 LTS and Spring 3.5, it moves beyond simple libraries to provide a deployable service with full conversation management, advanced orchestration, and enterprise-grade security.

🌟 Highlights

🚀 Standalone Service

Stop building the plumbing. uxopian-ai is a pre-packaged backend service that manages the complexities of LLM interactions. It is deployable immediately via Docker or as a Java 21 application, saving months of infrastructure setup.

🧠 The Goal System (Context-Aware Orchestration)

This is the heart of the framework. Instead of hardcoding prompts in your application code, you send a generic Goal (e.g., "compare"). The engine dynamically selects the correct Prompt based on context filters (Tenant, Document Type, User Role) using SpEL (Spring Expression Language).

🛡️ Native Multi-Tenancy

Built for the enterprise, every conversation, prompt, and statistic is strictly scoped to a Tenant ID. The architecture supports a BFF (Backend for Frontend) pattern, relying on security headers (X-User-TenantId) to ensure logical data separation.

✨ Key Features

🤖 LLM Interaction & Connectors

  • Provider Abstraction: Out-of-the-box support for major providers (OpenAI, Azure, Anthropic).
  • Parameter Precedence: A granular hierarchy for configuration. Parameters passed in the API call override Prompt defaults, which in turn override Global defaults.
  • Advanced Capabilities: Native support for Function Calling, Multi-modal inputs (Text + Base64 Images), and streaming responses.
  • MCP Client: Acts as a client for Multi-Content Platform (MCP) servers.

💬 Conversation Management

  • Persistent History: All conversations are stored in OpenSearch, managing the context window automatically.
  • Rich Requests: Requests track inputs, answers, and token usage.
  • User Feedback Loop: Built-in endpoints to tag responses as Good, Bad, or Neutral, enabling continuous improvement of prompt quality.

🎨 Frontend Components

  • Web Components: A set of lightweight, embeddable JS/CSS components to add a Chat UI to any web application instantly.
  • Thymeleaf Templating: Prompts and Image inputs support dynamic variable injection via Thymeleaf.

📊 Admin & Analytics Dashboard

A comprehensive administration panel to monitor the health and ROI of your AI stack:

  • ROI Tracking: Define "Time Saved" per prompt to calculate the estimated human hours saved.
  • Token Monitoring: Visualize input/output token consumption globally or per user.
  • Adoption Trends: Track daily requests, active users, and advanced feature adoption (e.g., Multi-modal usage).

🏗️ Architecture & Integrations

Technical Stack

  • Core: Java 21 LTS, Spring 3.5.
  • LLM Engine: Powered by Langchain4j.
  • Storage: OpenSearch (Vectors & Metadata).
  • API: Fully documented REST API with Swagger UI.

Ecosystem Integrations

v1.0.0 demonstrates the framework's flexibility through complex integrations with ARender and FlowerDocs.

These integrations were achieved seamlessly using standard JavaScript scripts and ScriptOperationHandlers, proving that uxopian-ai can be easily embedded into existing High Content Interfaces (HCI) and Content Services Platforms (CSP) without requiring heavy backend modifications.

👉 Ready to start? Check out the Installation Guide to deploy your first instance.