E.D.D.I Documentation

Multi-Agent Orchestration Middleware for Conversational AI — coordinate multiple AI agents, business systems, and conversation flows through configuration, not code.

Welcome to the official documentation for E.D.D.I (Enhanced Dialog Driven Interface) — a production-grade multi-agent orchestration middleware for conversational AI.

· License: Apache 2.0 · GitHub · Website

What Is EDDI?

EDDI coordinates between users, AI agents (LLMs), and business systems. It provides intelligent routing, conversation management, and API orchestration — all through versioned JSON configurations, not code.

Built with Java 25 and Quarkus. Ships as a Red Hat-certified Docker image. Supports MongoDB or PostgreSQL. Deploy on Docker, Kubernetes, or OpenShift.

Start Here

Guide

Time

Description

🚀 Getting Started

5 min

Install EDDI and run your first agent

⚡ Developer Quickstart

10 min

Build a complete agent step-by-step via REST API

🏗️ Architecture Overview

15 min

Understand the lifecycle pipeline and config model

🧩 Putting It All Together

20 min

Real-world hotel booking agent walkthrough

Key Capabilities

🤖 Multi-Agent Orchestration

12 LLM Providers — OpenAI, Anthropic, Google Gemini, Mistral AI, Azure OpenAI, Amazon Bedrock, Oracle GenAI, Vertex AI, Ollama, Jlama, Hugging Face, plus OpenAI-compatible endpoints
Group Conversations — Multi-agent debates, voting, shared artifacts, and standing teams across 7 discussion styles (Round Table, Peer Review, Devil's Advocate, Delphi, Debate, Task Force, Negotiation)
Managed Agents — Intent-based auto-routing with one conversation per user per intent
Model Cascading — Cost-optimized multi-model routing with confidence-based escalation
Platform Operator — Meta-agent that reads and operates the deployment, including creating other agents, with every write behind a human approval gate. Activate it at /manage/operator, or use the form-based wizard at /manage/agents/wizard — see Getting Started

🔗 Protocols & Interoperability

MCP Server (80+ tools) — Full EDDI control from Claude Desktop, IDE plugins, or any MCP client
MCP Client — Point an agent at somebody else's MCP server, including stdio servers via a bridge sidecar
A2A Protocol — Agent-to-Agent peer communication with skill discovery
Connections — One credential model for every outbound call: static keys, HTTP Basic, OAuth service accounts, and per-end-user OAuth
OpenAI-Compatible API — Deployed agents presented as OpenAI models for Open WebUI and OpenAI SDK clients
SSE Streaming — Token-by-token responses, including most tool-enabled turns (a single-chunk fallback applies to cascade agents, providers without a streaming builder, and a few other configurations), plus a live tool_call event so clients can show "Using {tool}…" while the turn is still running

🧠 Intelligence & Memory

LLM Integration — Connect any of 12 providers with agent mode and tool calling
RAG — 8 embedding providers, 6 vector stores, plus zero-infrastructure httpCall RAG
Persistent User Memory — Agents remember facts across conversations
Properties — Config-driven slot-filling and importance extraction

🔐 Enterprise Security

Secrets Vault — Envelope encryption (AES-256-GCM + PBKDF2) for API keys
Security — SSRF protection, sandboxed evaluation, Keycloak auth
Audit Ledger — Write-once trail with HMAC integrity for EU AI Act compliance
Human-in-the-Loop — Turn-level and per-tool-call approval gates with timeout policies, plus Slack and MCP approval surfaces

Agent Configuration

Build agent behavior by composing these extensions:

Extension

Purpose

Guide

Behavior Rules

Decision-making logic — IF conditions THEN actions