A team of AI employees designed to understand your business, perform specialized work, and deliver real business outcomes.
It enables AI to understand enterprise systems, enterprise data, enterprise knowledge and personal memory, then execute real business tasks through reusable Enterprise Skills.
Get Started → · Quick Deployment
Deploy Open Insight on your own server and start experiencing enterprise AI in minutes.
Enterprise AI starts with understanding the enterprise.
ONE-PI connects to an extensible set of expert agents, each equipped with Prompt, MCP, API, and Skill capabilities.
Most Data Agent systems are designed to answer questions.
Open Insight is designed to achieve business goals.
Enterprises don't measure conversations.
They measure outcomes.
Most AI Agents are designed for individual users.
Each user installs, configures and manages their own AI environment.
Open Insight takes a different approach.
The enterprise learns once. Everyone benefits.
Open Insight continuously learns from the enterprise — its data, systems, business knowledge, workflows, skills, and expertise. This knowledge is centrally managed, governed, and made available to employees across the organization.
Employees don't need to build their own AI from scratch. They simply use AI to accomplish their work to achieving business goals.
As employees use Open Insight, their interactions, feedback, and outcomes can further improve the AI.
It runs on enterprise infrastructure, not personal laptops. No installation, no setup, no maintenance burden on employees. It works 24/7 — monitoring systems, processing tasks, and collaborating with other AI agents across departments.
Open Core. Deploy anywhere.
| Repository | What it is | License |
|---|---|---|
| openinsight | One-command deployment & release entry for the whole platform | Open Core |
| arp | Agent Runtime Platform — the runtime that powers AI employees | Apache-2.0 |
| one-pi | Enterprise Agent Platform — expert agents, skills, and orchestration | Apache-2.0 |
The enterprise edition (DMP) is available under a commercial license, activated by uploading a License File. See LICENSE-COMMERCIAL.md.
One stack, one roadmap.
| Component | Shipped | Next |
|---|---|---|
| ARP — where AI employees run | Multi-provider chat · agents & MCP · shared skills · self-hosted deploy | Deeper ONE-PI integration · community agent & MCP templates · local browser extension — client-side reach beyond the browser/server split |
| ONE-PI — how AI employees think | Virtual experts · shared skill repository · OpenAI-compatible agent API | A2A collaboration · more experts |
| DMP — what enterprises get | Four learning engines · governed enterprise core (commercial) | Deeper enterprise knowledge base · more enterprise connectors · deeper governance |
A2A, redefined. Multi-agent systems let one framework orchestrate many agents. A2A in Open Insight is different: every person commands their own AI, and the collaboration that used to happen person-to-person now happens agent-to-agent — with humans approving what matters.
The core is open source. The enterprise edition builds on it.
Deploy Open Insight on your own infrastructure and experience the Enterprise Agent Operating System.