EDDI

Overview

EDDI (Enhanced Dialog Driven Interface) is a config-driven engine that transforms JSON into production-ready AI agents. Built on Java 25 and Quarkus, it provides multi-agent orchestration with intelligent routing, persistent memory, and API integration. The platform supports 12+ LLM providers, compliance frameworks (EU AI Act, GDPR, HIPAA), and includes native MCP support for seamless tool integration. With features like group conversations, agent capability matching, and envelope-encrypted secrets, EDDI offers enterprise-grade security and governance for AI systems.

Try asking AI

After installing, here are 5 things you can ask your AI assistant:

• agent_father_create

  Create new AI agents through conversation using the meta-agent Agent Father

  Try:Create an agent that can help me analyze financial reportsTry:Build a customer service agent for my e-commerce site
• group_conversation

  Facilitate multi-agent debates with 5 built-in discussion styles

  Try:Start a peer review of this new marketing strategyTry:Run a round table discussion about product improvements
• conversation_routing

  Route conversations to different agents based on context, rules, and intent

  Try:Route this technical support question to our specialized agentTry:Forward this customer feedback to the appropriate team
• persistent_memory

  Store and retrieve user memory across conversations

  Try:Remember that I prefer to communicate via email for important updatesTry:What was the name of the client I spoke to yesterday?
• llm_task_execution

  Execute LLM tasks with configurable parameters, RAG, and budget settings

  Try:Generate a summary of this article using Claude 3Try:Translate this document to French with GPT-4
• schedule_execution

  Run scheduled tasks with cron jobs, heartbeats, and retry logic

  Try:Set up a daily report to be generated at 9 AMTry:Create a heartbeat check for our critical services
• secrets_management

  Manage envelope-encrypted secrets with rotation and access control

  Try:Store the API key for our payment processorTry:Rotate the database credentials for production
• tenant_quota_monitoring

  Monitor rate limits, cost budgets, and usage per tenant

  Try:Check the current usage for our team this monthTry:Set up a budget alert when we reach 80% of our quota
• openapi_agent_generator

  Create API-calling agents from OpenAPI specifications

  Try:Create an agent for our HR API documentationTry:Build a customer service agent from our support API
• agent_capability_matching

  Discover and route to agents by skill, confidence score, and attributes

  Try:Find an agent who can help with JavaScript debuggingTry:Connect me with an agent specializing in compliance audits
• real_time_logging

  Access live log streams with cost tracking and token counts

  Try:Show me the logs for the last hourTry:Track the token usage for this conversation
• slack_integration

  Deploy agents to Slack channels and run conversations directly in threads

  Try:Deploy our customer support agent to #helpdeskTry:Start a discussion in #planning about the new feature

Comparable tools

Installation

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

# Windows (PowerShell)
Invoke-WebRequest -UseBasicParsing -Uri "https://raw.githubusercontent.com/labsai/EDDI/main/install.ps1" -OutFile "install.ps1"
Unblock-File .\install.ps1
.\install.ps1

{
  "mcpServers": {
    "eddi": {
      "command": "docker",
      "args": ["run", "--rm", "-i", "labsai/eddi", "mcp"]
    }
  }
}

FAQ

What makes EDDI different from other multi-agent frameworks?

EDDI is a production-grade middleware built on Java 25 with Quarkus, offering deterministic orchestration of AI agents. Unlike Python/Node frameworks, it provides versioned JSON configurations, no dynamic code execution, built-in compliance (EU AI Act, GDPR, HIPAA), and enterprise-grade security with envelope-encrypted secrets.

How does EDDI integrate with MCP?

EDDI has native MCP support, allowing it to expose tools and resources through the Model Context Protocol. It can be configured as an MCP server in Claude Desktop or other compatible clients using Docker containers, enabling seamless integration with other AI systems.

On Hacker News

Recent discussion from the developer community.

• Show HN: EDDI – Multi-agent AI engine where agent logic lives in JSON, not code▲ 2
  Story by ginccc · 2026-04-16

Compare EDDI with