labsai/eddi repository overview

⁠🚀 EDDI — Multi-Agent Orchestration Engine for Conversational AI

EDDI (Enhanced Dialog Driven Interface) is a production-grade, config-driven multi-agent orchestration middleware. Define agent behavior in JSON configurations — not code. Built with Java 25 and Quarkus. Ships as a single Docker image.

🌐 Website: eddi.labs.ai⁠ · 📖 Docs: docs.labs.ai⁠ · 💻 GitHub: labsai/EDDI⁠ · 📜 License: Apache 2.0

⁠⚡ Quick Start (Recommended)

The fastest way to get started — an interactive installer that sets up EDDI + database + a starter agent:

# Linux / macOS / WSL2
curl -fsSL https://raw.githubusercontent.com/labsai/EDDI/main/install.sh | bash

# Windows (PowerShell)
iwr -useb https://raw.githubusercontent.com/labsai/EDDI/main/install.ps1 | iex

The wizard auto-generates a vault encryption key and deploys the Agent Father — a meta-agent that creates other agents through conversation.

⁠🐳 Docker Compose

Download docker-compose.yml⁠ from GitHub, then:

docker compose up -d

This starts EDDI v6 + MongoDB, fully configured and ready to go.

Additional overlays (combine as needed):

# PostgreSQL instead of MongoDB
EDDI_DATASTORE_TYPE=postgres docker compose -f docker-compose.yml -f docker-compose.postgres.yml up -d

# With Keycloak authentication
docker compose -f docker-compose.yml -f docker-compose.auth.yml up -d

# With Prometheus + Grafana monitoring
docker compose -f docker-compose.yml -f docker-compose.monitoring.yml up -d

⁠🛠️ Manual Container Launch

If you prefer to start containers individually:

# 1. Create a shared network
docker network create eddi-network

# 2. Start MongoDB
docker run --name mongodb --network eddi-network -d mongo:6.0

# 3. Start EDDI
docker run --name eddi --network eddi-network -p 7070:7070 -p 7443:7443 -d labsai/eddi:latest

⁠✅ Verifying Your Installation

Once EDDI is running, access these endpoints locally:

⁠✨ Key Features

• 12 LLM Providers — OpenAI, Anthropic, Google Gemini, Mistral, Azure OpenAI, Bedrock, Ollama, and more
• Multi-Agent Orchestration — Intelligent routing, group conversations (5 discussion styles), nested groups
• MCP Server (42 tools) — Control EDDI from Claude Desktop, Cursor, or any MCP client
• A2A Protocol — Agent-to-Agent peer communication and skill discovery
• Persistent User Memory — Cross-conversation fact retention with dream consolidation
• RAG — 7 embedding providers, 5 vector stores, httpCall RAG
• Smart Model Cascading — Try cheap models first, escalate only when confidence is low
• Enterprise Security — AES-256 secrets vault, SSRF protection, agent signing, no dynamic code execution
• Compliance — GDPR, HIPAA, EU AI Act, CCPA infrastructure built-in
• Config-Driven — Agent behavior in versioned JSON — update without redeployment
• DB-Agnostic — MongoDB or PostgreSQL — switch with one environment variable
• Quarkus SDK — quarkus-eddi⁠ extension with Dev Services and @Inject EddiClient

⁠🔧 Configuration

EDDI is configured via environment variables:

See the full documentation⁠ for all configuration options.

⁠📚 Resources

• 🌐 Website: eddi.labs.ai⁠
• 📖 Documentation: docs.labs.ai⁠
• 💻 Source Code: github.com/labsai/EDDI⁠
• 🧩 Quarkus SDK: github.com/quarkiverse/quarkus-eddi⁠
• 🐳 Docker Hub: hub.docker.com/r/labsai/eddi⁠
• 🤝 Contributing: CONTRIBUTING.md⁠
• 🔒 Security Policy: SECURITY.md⁠