Ontheia — Your Data. Your AI. Your Rules.

# Ontheia > Your Data. Your AI. Your Rules. — Ontheia is the self-hosted AI agent platform for organizations that cannot afford to send their data to external cloud services. Ontheia solves a concrete problem: every request to a cloud AI (ChatGPT, Gemini, Claude via third-party services) sends your data to foreign servers — uncontrolled and often without a clear legal basis. Ontheia is the alternative: the same AI capabilities, but running exclusively on your own infrastructure. Your data never leaves your servers. Ontheia runs on Docker, deploys in under 30 minutes, and connects any AI model (Anthropic Claude, OpenAI GPT, Google Gemini, local models via Ollama) to your tools and data via the MCP protocol. ## The Problem — why Ontheia exists **Data leaking through the back door:** Every request to a cloud AI sends your data to foreign servers — uncontrolled, opaque. **Vendor lock-in:** One provider change, one price hike, one updated ToS — and your AI stack collapses. **Compliance risk:** Personal data in US cloud services is legally precarious in the EU. That costs nerves — and potentially money. ## What Ontheia is — Core Capabilities **Self-hosted:** Runs on your own infrastructure. Docker-based, easy to deploy. On-premise or VPS — your choice. **Multi-provider:** Claude, ChatGPT, Grok, Gemini, Ollama — you choose the model. No lock-in. Define per agent which model sees which data. **Multi-User:** Role-based access control, teams, and projects — Ontheia is built for organizations from the ground up. **Workflow Engine:** Visual automation pipelines. Complex agent logic without code. Agents execute multi-step tasks autonomously. **Long-term Memory (RAG):** Built-in RAG with pgvector. Your agents remember and access your company knowledge. Documents, manuals, past projects — all searchable. **Agent-to-Agent:** Agents communicate directly with each other. Complex tasks can be split and coordinated between agents. ## How it works O