Get Started with EDDI

Prerequisites

• Docker (recommended) or Java 25+
• An LLM provider API key (OpenAI, Anthropic, Google Gemini, or a local Ollama instance)

1. Install & Start EDDI

The fastest way to get EDDI running is the one-command installer. It sets up EDDI + your choice of database via Docker Compose, deploys a starter agent, and walks you through configuration:

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

iwr -useb https://raw.githubusercontent.com/labsai/EDDI/main/install.ps1 | iex

docker compose up

EDDI will be available at http://localhost:7070.

2. Connect via MCP

If you use Claude Desktop or any MCP-compatible client, add EDDI as an MCP server:

{
  "mcpServers": {
    "eddi": {
      "url": "http://localhost:7070/mcp"
    }
  }
}

Now you can interact with EDDI's 42 MCP tools directly from your AI assistant.

3. Store Your API Key

Before creating an agent, securely store your LLM provider API key in EDDI's Secrets Vault (AES-256-GCM encrypted):

curl -X PUT http://localhost:7070/secretstore/secrets/default/my-anthropic-key \
  -H "Content-Type: application/json" \
  -d '{"value": "sk-ant-your-actual-key", "description": "Anthropic API key"}'

💡 Tip: The ${eddivault:my-anthropic-key} syntax references EDDI's built-in Secrets Vault. The vault master key is auto-generated by the installer. For quick testing, you can also pass API keys directly: apiKey: "sk-ant-...".

4. Create Your First Agent

Use setup_agent to create a fully working agent in one call — via MCP or REST API:

setup_agent(
  agentName: "My first agent",
  systemPrompt: "You are a helpful assistant that answers questions clearly.",
  provider: "anthropic",
  model: "claude-sonnet-4-6",
  apiKey: "${eddivault:my-anthropic-key}"
)

curl -X POST http://localhost:7070/administration/agents/setup \
  -H "Content-Type: application/json" \
  -d '{
    "agentName": "My first agent",
    "systemPrompt": "You are a helpful assistant that answers questions clearly.",
    "provider": "anthropic",
    "model": "claude-sonnet-4-6",
    "apiKey": "${eddivault:my-anthropic-key}"
  }'

This creates the rules, LLM config, workflow, agent, and deploys it — all in one step.

5. Chat with Your Agent

chat_with_agent(agentId: "<your-agent-id>", message: "Hello! What can you do?")

# Start a conversation and send a message
curl -X POST http://localhost:7070/agents/<your-agent-id>/start \
  -H "Content-Type: application/json" \
  -d '{"input": "Hello! What can you do?"}'

Installer Options

The install script supports flags for automated and customized setups:

• --defaults — All defaults, no prompts
• --db=postgres — Use PostgreSQL instead of MongoDB
• --with-auth — Enable Keycloak authentication
• --full — Everything enabled (database + auth + monitoring)
• --local — Build Docker image from local source (for contributors)

Updating EDDI

The installer creates an eddi CLI wrapper. To pull the latest image and restart:

eddi update

If the CLI is not available, run from your install directory (~/.eddi):

cd ~/.eddi
docker compose --env-file .env -f docker-compose.yml pull
docker compose --env-file .env -f docker-compose.yml up -d

Kubernetes Deployment

Deploy to Kubernetes with a single command:

kubectl apply -f https://raw.githubusercontent.com/labsai/EDDI/main/k8s/quickstart.yaml

Kustomize overlays and Helm charts are also available for MongoDB, PostgreSQL, auth, monitoring, and production hardening (HPA, PDB, NetworkPolicy). See the Kubernetes Guide.