What is Open Enthrium?

Open Enthrium is a self-hosted, open-source AI Intelligence Platform and AI Agent Runtime for building and running autonomous AI agents without writing code. Define multi-step AI workflows in plain YAML and connect to enterprise systems—including databases, APIs, cloud storage, messaging, SSH, and IoT—without custom integration code.

Is Open Enthrium free?

Yes. The Community Edition is free forever under Apache-2.0. You can self-host it on any infrastructure at no cost. A Commercial Edition is available for organizations needing SSO, DLP, white-label, multi-tenant workspaces, and dedicated support.

How do I install Open Enthrium?

Run: docker pull jumbocoder/open-enthrium-ai-platform:latest then start it with JWT_SECRET, SUPER_ADMIN_EMAIL, and SUPER_ADMIN_PASSWORD environment variables. The platform is ready in minutes with no build step required.

What LLM providers does Open Enthrium support?

Open Enthrium supports 17+ LLM providers including OpenAI, Anthropic Claude, Azure OpenAI, Ollama, LM Studio, Together AI, Google Gemini, and more. Switch providers anytime from the admin panel without redeploying.

OE Runtime is a standalone binary that executes any Open Enthrium agent YAML file locally — on Windows, Linux, or macOS. No Docker, no database, no Node.js required. Download the binary, point it at a YAML agent file and a config, and run. It supports 20 capability categories including SQL databases, SSH, REST APIs, email, Slack, image generation, web search, and more.

What is the difference between the Community and Commercial editions?

The Community Edition is free and open-source. The Commercial Edition adds SSO/SAML, Data Loss Prevention (DLP), white-label branding, multi-tenant workspaces, visual agent builder, REST API with Swagger, cross-workspace collaboration, enterprise infrastructure (PostgreSQL, S3, Kubernetes), and dedicated support with SLAs.

Open Source & Free to Self-Host

Run AI Anywhere

Run your SKILL.md via npx, binary, API server, or web app.
Self-hosted. Any LLM. Your data. Apache-2.0.

Apache-2.0 Licensed Self-Hosted No LangChain. No Python. No code. 17+ LLM Providers 45+ Connector Categories Agent Chains

Get Started

            # run a skill instantly

            npx -y @openenthrium/oe-runtime@latest agent.yaml

GitHub Downloads →

Quick Start

          hello-world/

            ├── SKILL.md← skill instructions

            ├── agent.yaml← entry point

            └── oe-config.json← LLM + connectors

No install needed — always runs the latest version

SKILL.md

---
name: hello-world
description: A simple greeting skill
license: Apache-2.0
metadata:
author: Open Enthrium
version: "1.0"
---

## Step 1: Greet
Introduce yourself as an AI
assistant powered by Open Enthrium.

## Step 2: Today's Briefing
Today is Aug 25. Share one notable
tech event from this date in history.

agent.yaml

name: Hello World
description: A greeting skill
skills:
- path: ./
trigger_type: auto

oe-config.json

{
"llm": {
  "provider": "openai",
  "model": "gpt-4o",
  "apiKey": "sk-..."
},
"server": {
  "enabled": false,
  "port": 3333,
  "apiKey": "your-secret-key"
},
"connectors": []
}

# option 1 — no install, always latest
npx -y @openenthrium/oe-runtime@latest agent.yaml
# option 2 — install once globally, run anywhere
npm install -g @openenthrium/oe-runtime@latest
oe-runtime agent.yaml

Download once, run offline — no Node.js required

SKILL.md

agent.yaml

name: Hello World
description: A greeting skill
skills:
- path: ./
trigger_type: auto

oe-config.json

{
"llm": {
  "provider": "openai",
  "model": "gpt-4o",
  "apiKey": "sk-..."
},
"server": {
  "enabled": false,
  "port": 3333,
  "apiKey": "your-secret-key"
},
"connectors": []
}

⊞ Windows (.exe) macOS 🐧 Linux

# Windows
oe-runtime-win.exe agent.yaml
# macOS / Linux
chmod +x oe-runtime-macos && ./oe-runtime-macos agent.yaml

Expose a local HTTP endpoint — call from any app, webhook, or automation pipeline

SKILL.md

agent.yaml

name: Hello World
description: A greeting skill
skills:
- path: ./
trigger_type: auto

oe-config.json

{
"llm": {
  "provider": "openai",
  "model": "gpt-4o",
  "apiKey": "sk-..."
},
"server": {
  "enabled": true,
  "port": 3333,
  "apiKey": "your-secret-key"
},
"connectors": []
}

npx -y @openenthrium/oe-runtime@latest agent.yaml --serve
# call from any app or pipeline
curl -X POST http://localhost:3333/run-file \
  -H "Content-Type: application/json" \
  -H "X-API-Key: your-secret-key" \
  -d '{"file": "hello-world/agent.yaml"}'

Embed agents directly in your Node.js app — no server, no HTTP round-trip

SKILL.md

agent.yaml

name: Hello World
description: A greeting skill
skills:
- path: ./
trigger_type: auto

oe-config.json

{
"llm": {
  "provider": "openai",
  "model": "gpt-4o",
  "apiKey": "sk-..."
},
"server": {
  "enabled": false,
  "port": 3333,
  "apiKey": "your-secret-key"
},
"connectors": []
}

npm install @openenthrium/oe-runtime-sdk
// in your app
const { runAgent } = require("@openenthrium/oe-runtime-sdk");
const result = await runAgent(
  "hello-world/agent.yaml",
  "hello-world/oe-config.json",
  {},
  { onToolCall: name => console.log(`🔧 ${name}`) }
);
console.log(result.output);

Skip SKILL.md entirely — write steps directly in agent.yaml. One file, zero boilerplate.

agent.yaml

name: Hello World
instructions: You are a helpful assistant.
steps:
- name: Greet
  content: Say hello and today's date.

- name: Tech Fact
  content: Share one notable tech
          moment from history.

- name: Sign Off
  content: Close with a motivational
          quote for the day.

oe-config.json

{
"llm": {
  "provider": "openai",
  "model": "gpt-4o",
  "apiKey": "sk-..."
},
"connectors": []
}

# just two files — no SKILL.md needed
npx -y @openenthrium/oe-runtime@latest agent.yaml

Full platform — upload your skill folder, run from the browser, schedule, and manage access

SKILL.md

agent.yaml

name: Hello World
description: A greeting skill
skills:
- path: ./
trigger_type: auto

oe-config.json

{
"llm": {
  "provider": "openai",
  "model": "gpt-4o",
  "apiKey": "sk-..."
},
"server": {
  "enabled": false,
  "port": 3333,
  "apiKey": "your-secret-key"
},
"connectors": []
}

# deploy OE Platform
docker run -d --name open-enthrium \
  -p 3000:3000 \
  -p 3001:3001 \
  -v open-enthrium-storage:/app/server/storage \
  -e JWT_SECRET=your-secret \
  -e SUPER_ADMIN_EMAIL=admin@company.com \
  -e SUPER_ADMIN_PASSWORD=your-password \
  -e LICENSE_TYPE=community \
  jumbocoder/open-enthrium-ai-platform:latest
# open http://localhost:3000 — log in with your super admin credentials

Use your agents as MCP tools in Claude Code, Cursor, or Windsurf — invoke via natural language, no server needed

SKILL.md

agent.yaml

name: Hello World
description: A greeting skill
skills:
- path: ./
trigger_type: auto

.mcp.json

{
"mcpServers": {
  "oe-mcp": {
   "type": "stdio",
   "command": "npx",
   "args": [
    "-y",
    "@openenthrium/oe-mcp",
    "--stdio",
    "/path/to/oe-mcp.json"
   ]
  }
}
}

# ask in Claude Code / Cursor / Windsurf chat:
"Run my agent at /home/user/hello-world/agent.yaml"
# Claude Desktop — paste same JSON into claude_desktop_config.json
# macOS: ~/Library/Application Support/Claude/claude_desktop_config.json
# Windows: %APPDATA%\Claude\claude_desktop_config.json

Claude Code reads SKILL.md and calls your connectors as tools — no oe-config.json, no API key, no cost beyond your Claude Code subscription

SKILL.md

oe-mcp.json

{
"connectors": [
  {
   "name": "my-slack",
   "type": "slack",
   "botToken": "xoxb-YOUR-TOKEN"
  },
  {
   "name": "my-postgres",
   "type": "postgresql",
   "host": "localhost",
   "port": 5432,
   "database": "mydb",
   "user": "postgres",
   "password": "secret"
  }
]
}

.mcp.json

{
"mcpServers": {
  "oe-mcp": {
   "type": "stdio",
   "command": "npx",
   "args": [
    "-y",
    "@openenthrium/oe-mcp",
    "--stdio",
    "/path/to/oe-mcp.json"
   ]
  }
}
}

# tell Claude Code / Cursor / Windsurf — no API key needed:
"Read my SKILL.md at /home/user/hello-world/SKILL.md and follow the steps"
# Claude Code is the LLM — uses your Claude Code / Cursor subscription
# connector tools in oe-mcp.json are exposed via oe-mcp as MCP tools