labsai/EDDI

Config-driven engine that turns JSON into production-grade AI agents. Multi-agent orchestration, 19+ LLM providers, MCP/A2A protocols, RAG, persistent memory, and enterprise compliance (EU AI Act, GDPR, HIPAA). Built on Quarkus.

E.D.D.I – Multi‑Agent Orchestration Middleware for Conversational AI

What it is – E.D.D.I (Enhanced Dialog‑Driven Interface) is a production‑ready, config‑driven platform that lets you run, manage and scale many AI agents that talk to users and to each other. It sits between a user (or a business system) and large‑language‑model providers, handling routing, memory, security and orchestration without writing code.

Key ideas

  • No‑code configuration – agents, workflows and group‑discussions are described in JSON files. Changing a file updates behaviour; you never need to rebuild or redeploy the Java service.
  • Java 25 + Quarkus – the engine runs on a modern, high‑performance Java runtime that uses virtual threads for true parallelism, giving better concurrency than typical Python/Node frameworks.
  • Docker‑first delivery – a single Docker image (Red‑Hat certified) contains the server, a web‑based Manager dashboard and a chat UI. The install.sh script can spin up the whole stack (database, optional Keycloak auth, monitoring, local LLM) with one command.
  • Enterprise‑grade features – built‑in GDPR/HIPAA compliance, immutable audit logs, encrypted secret vault, tenant quotas, and support for OpenSSF best‑practice badges.

Main capabilities

Area What E.D.D.I provides
Multi‑agent orchestration Intelligent routing, group‑conversation styles (round‑table, debate, negotiation, etc.), dynamic recruitment of new agents, voting, auctions, and persistent team memory.
LLM integration Connects to 19 providers (OpenAI, Anthropic, Gemini, Bedrock, Ollama, etc.). Supports model cascading, JSON‑only responses and tool‑calling with per‑tool caching.
Interoperability Implements open protocols such as MCP (Model Context Protocol), A2A (Agent‑to‑Agent), OpenAI Chat Completions, Slack, OpenAPI and OAuth 2.0.
Security & compliance Envelope‑encrypted vault, SSRF protection, HMAC‑SHA256 immutable ledger, GDPR/HIPAA/AI‑Act ready controls, role‑based access, tenant‑level rate‑limits and budgets.
Observability Real‑time logs, token‑/cost‑tracking, Prometheus + Grafana metrics, dashboards for conversations, agents and quotas.
Deployment options Docker‑Compose (with overlays for Postgres, Keycloak, monitoring, NATS, local LLM, vector store, etc.) and ready‑to‑run on Kubernetes/OpenShift.

Typical workflow

  1. Run install.sh (or use the provided Docker‑Compose files) – the script creates a eddi CLI wrapper.
  2. Open http://localhost:7070 → choose Manager to define agents via the form‑based wizard or by uploading JSON configs.
  3. Deploy an agent that points to a LLM endpoint (e.g., OpenAI or a local Ollama model).
  4. Users start a chat at /chat or Slack threads; E.D.D.I routes the request to the appropriate agent or group.
  5. The platform records every step, updates memory, enforces quotas and, if configured, escalates to a higher‑tier model.

Who it’s for

  • Enterprises that need a governed, auditable AI‑agent layer rather than ad‑hoc scripts.
  • Teams building customer‑service bots, internal assistants, or multi‑agent decision‑support systems that must comply with regulatory standards.
  • Ops teams that prefer Docker/Kubernetes deployment and want to avoid Python‑centric runtimes.

Where to learn more

  • Official website: https://eddi.labs.ai/
  • Docs: https://docs.labs.ai/
  • Quick‑start installer script in the repo (install.sh / install.ps1).
  • Release notes and upgrade guides are in docs/ (e.g., upgrading-from-6.4.md).

All information above is taken directly from the repository’s README; no additional features have been inferred.

Related

  • Project
    VRSEN/agency-swarm

    A framework for building multi-agent AI applications that organizes agent collaboration using real-world organizational structures and directional communication flows.

  • Project
    zts212653/clowder-ai

    Clowder AI is a platform layer that turns isolated AI coding agents into a collaborative team, giving them persistent identities, shared memory, cross-model review, and agent-to-agent messaging so you don't have to manually route copy-paste context.

  • Project
    landingbj/LinkMind

    An enterprise-grade multimodal AI middleware that provides a unified API layer for managing multiple model providers, RAG, and agent runtimes with centralized routing and governance.

  • Project
    evalstate/fast-agent

    A flexible framework for building multimodal AI agents and workflows with deep Model Context Protocol (MCP) integration and advanced orchestration patterns.

  • Project
    steve02081504/fount

    fount is a programmable, modular AI agent runtime platform that lets you create and customise AI characters for productivity and companionship, with features like live code execution, P2P collaboration, and broad integration. It's built on Deno/Express and uses a part-based architecture. The project is currently in a development pause while the author works on a custom language.