unified-ui – Open Source AI Integration Platform
Universal integration platform for AI agents – one interface for Foundry, N8N, LangGraph and custom solutions with central tracing and Enterprise RBAC.
About the Project
Companies deploy AI agents across various platforms – Microsoft Foundry, N8N workflows, LangGraph or custom REST APIs. This leads to fragmented experiences: each platform has its own chat interface, custom agents often have no UI at all, and monitoring data is scattered across different systems. Additionally, the rapid evolution of agent frameworks means what is state-of-the-art today can be legacy tomorrow.
The Solution
unified-ui solves this problem through consistent decoupling of agent framework and user experience. The platform provides a unified chat interface for all agents, regardless of their technical implementation. This enables seamless transitions between frameworks without disrupting end users.
Architecture
The architecture consists of five loosely coupled services: A React frontend with Mantine UI, a FastAPI-based Platform Service for tenant, permission and configuration management, a Go-based Agent Service for execution and streaming of agent responses, and a Python SDK for integrating external agents. Communication occurs via Server-Sent Events (SSE) for real-time streaming.
Enterprise RBAC
The RBAC system supports multi-tenant organizations with granular permissions at tenant and resource level. Various identity providers are supported (Entra ID, Google, AWS, LDAP and more). Credentials are securely stored in Azure Key Vault or HashiCorp Vault.
Widget System
Particularly noteworthy is the widget system: Chat widgets like forms or custom iFrames can be embedded directly into the chat experience. This allows data collection, approval processes or interactive dashboards to be integrated into conversations – the chat becomes a full-fledged work environment.
Unified Tracing
A central feature is unified tracing: Traces from all agents/workflows – whether Foundry, N8N or custom solutions – are centrally collected, made semantically searchable and displayed in a visual hierarchy. The SDK includes callbacks for LangChain and LangGraph that automatically stream traces to the platform. Traces can also be fed into the platform externally via the unified-ui-sdk.
Highlights
- Framework-agnostic integration of AI agents (Foundry, N8N, LangGraph, Custom REST)
- Central tracing visualization with canvas and hierarchy views
- Widget system for forms, iFrames and interactive components
- Enterprise RBAC with various identity providers (Entra ID, Google, AWS, LDAP)
Challenges
Abstracting different agent protocols into a unified streaming format. Performant tracing at high message volume. Consistent developer experience across 5 services in 3 languages.
Results
Open-source platform at github.com/unified-ui. 1900+ tests in Platform Service with 80%+ coverage. Production-ready for Azure and on-premises. Future-proof architecture for seamless framework changes.
Interested?
Are you interested in a similar project or want to learn more? Get in touch – we look forward to hearing from you!
Get in touchRelated Projects
Azure AI Landing Zone Framework
Enterprise framework for declarative provisioning of AI infrastructure on Azure – from Microsoft Foundry to AI Search to private networks.
AI Research Application (Healthcare)
AI-powered research chatbot for evaluating medical and operational equipment based on internal and public data sources.