# Open Enterprise

> An open-source enterprise AI ecosystem. Build AI assistants, deploy autonomous agents, connect 2,500+ enterprise connectors, and govern AI securely — all self-hosted on your own infrastructure. Free under Apache-2.0.

## What is Open Enterprise?

Open Enterprise is a lightweight, self-hosted enterprise AI platform that ships as a single Docker container. It provides everything an organization needs to deploy AI internally: workspaces, agents, connectors, knowledge bases, and governance — without sending data to a third-party SaaS.

- License: Apache-2.0 (Community Edition), Commercial License (Commercial Edition)
- Deployment: Single Docker container, Docker Compose, Kubernetes
- Current version: 1.3.1
- GitHub: https://github.com/openenterprise-info/open-enterprise-community
- Docker Hub: https://hub.docker.com/r/openenterprise/open-enterprise-community

## Community Edition (Free)

Available at: https://www.openenterprise.info

Features:
- AI Assistants & Chat
- AI Agents (YAML-based, no code required)
- Knowledge Base with RAG (24+ file formats)
- 2,500+ connectors across 39 categories (limited)
- Website Chat Widget (Embed)
- Role-Based Access Control (RBAC)
- Enterprise Dashboard
- Vector Database support: LanceDB, Pinecone, Weaviate, Qdrant, Chroma, Milvus, Zilliz, Astra DB
- 17+ LLM providers: OpenAI, Anthropic Claude, Azure OpenAI, Ollama, LM Studio, Together AI, Google Gemini, and more
- CLI (limited)
- Marketplace (limited)
- Single Docker deployment
- SQLite database

## Commercial Edition

Contact: team@openenterprise.info

Additional features over Community Edition:
- Visual Agent Builder (no-code)
- Multi-Tenant Workspaces (full tenant isolation)
- REST API & OpenAPI / Swagger documentation
- Cross-Workspace Knowledge, Agent, and Connector Sharing
- SSO / SAML (Okta, Azure AD, Google Workspace, any SAML 2.0 provider)
- Data Loss Prevention (DLP)
- Audit Logs & Activity History
- Token Usage & Cost Analytics
- Super Admin & Tenant Management
- Custom Branding (White-Label — logo, product name, theme)
- Unlimited Connectors & Marketplace
- Enterprise Infrastructure: PostgreSQL, S3/MinIO, BullMQ + Redis, Kubernetes
- Priority Support, SLA-backed Support, Long-Term Support (LTS)
- Deployment Assistance, Training & Enablement, Professional Services

## OE Runtime

OE Runtime is a standalone, open-source agent execution engine — a single binary that runs YAML-defined agents locally with no server, no database, and no UI required. It is separate from the Docker platform and is designed for developers, power users, and CI/CD pipelines that want to run agents from the command line or via a lightweight HTTP API.

- License: Apache-2.0
- Platforms: Windows, macOS, Linux (single binary, no dependencies)
- Runtime page: https://www.openenterprise.info/runtime.html
- GitHub Releases: https://github.com/openenterprise-info/open-enterprise-community/releases/latest

### How it works

1. Write an `agent.yaml` file (name, description, instructions, steps, connectors, optional params)
2. Create an `oe-config.json` file with your LLM provider, API keys, and connector credentials
3. Run the binary: `./oe-runtime run agent.yaml --config oe-config.json`

### Supported fields in agent.yaml

- `name` — agent display name
- `description` — what the agent does
- `instructions` — system prompt for the LLM
- `steps` — ordered list of tasks (name + content); the LLM executes each step in sequence using available tools
- `connectors` — list of named connectors (connection_name + connection_type) the agent can call
- `params` — optional named parameters with defaults; substituted into instructions and step content via `{{param_name}}` syntax using `--param key=value` at runtime
- `maxRounds` — optional per-agent tool-call limit (overrides workspace default, overrides hardcoded fallback of 25)

### Downloads

- Windows: https://github.com/openenterprise-info/open-enterprise-community/releases/latest/download/oe-runtime-win.exe
- macOS: https://github.com/openenterprise-info/open-enterprise-community/releases/latest/download/oe-runtime-macos
- Linux: https://github.com/openenterprise-info/open-enterprise-community/releases/latest/download/oe-runtime-linux
- Sample agents (20 pre-built YAML agents): https://github.com/openenterprise-info/open-enterprise-community/releases/latest/download/oe-runtime-samples.zip
- Postman collection: https://github.com/openenterprise-info/open-enterprise-community/releases/latest/download/oe-runtime.postman_collection.json

### Server mode

Set `server.enabled: true` in `oe-config.json` to expose a REST API on port 3333. Endpoints:
- `GET /health` — liveness check
- `POST /run` — run an agent by passing YAML inline
- `POST /run-file` — run an agent from a YAML file path on the server's disk

### Pre-built agent categories (20 sample agents)

blockchain-web3, cloud-drives, directory-identity, email, file-storage, graphql, image-generation, iot-messaging, message-queues, music-generation, nosql-cache, ocr-vision, productivity-crm, rest-api, speech-audio, sql-databases, ssh, team-messaging, video-generation, web-search

## Quick Start

```bash
docker run -d \
  -p 3001:3001 \
  -e JWT_SECRET=your-secret \
  -e SUPER_ADMIN_EMAIL=admin@yourcompany.com \
  -e SUPER_ADMIN_PASSWORD=your-password \
  -v ./data:/app/server/storage \
  openenterprise/open-enterprise-community:latest
```

## Links

- Website: https://www.openenterprise.info
- Blog: https://www.openenterprise.info/blog/
- GitHub: https://github.com/openenterprise-info/open-enterprise-community
- Docker Hub: https://hub.docker.com/r/openenterprise/open-enterprise-community
- Discord Community: https://discord.gg/vWsZ24Msn
- Commercial Enquiries: team@openenterprise.info

## Blog Topics

Open Enterprise publishes educational content on enterprise AI, self-hosted AI, AI agents, RAG, workflow automation, and industry-specific AI use cases. See https://www.openenterprise.info/blog/ for all posts.