About
OpenAgentic is Chat, Flows, and a code surface in one self-hosted platform — AI that operates your infrastructure through real tools, on your hardware, against the models you choose, behind a human-approval gate on every write. One command and it's yours.
curl -sSL https://install.openagentics.io | bashThree surfaces — Chat, Flows, and code — over a set of pre-wired MCP tools, a multi-provider model router, human-in-the-loop gates on anything that mutates, and a tamper-evident audit trail. It runs on Docker Compose or Kubernetes, fully air-gapped if you want, against local Ollama or any cloud provider you authorize.
OpenAgentic wasn't born in a demo. It's the distillation of years spent operating real agents in some of the most demanding, tightly-regulated environments there are — where nothing leaves the boundary, every action has to name a human, and "it works on my machine" isn't a standard anyone accepts. We took what that taught us and made it something anyone can run.
OpenAgentic is the self-hosted core, free to run forever on your own infrastructure. AgenticWork is the enterprise tier on the same lineage — managed SaaS when an org wants it, or a white-glove engagement — adding multi-tenancy, SSO, DLP, and the operational commitments a regulated team needs. Same architecture, two ownership postures.
Every mutating action passes a human-approval gate; the audit chain is tamper-evident end to end; data and telemetry stay inside your boundary. Per-user identity — running each action as the signed-in human via SSO and on-behalf-of token exchange — is part of the Gnomus enterprise tier. The architecture is designed to meet ATO in Azure, AWS, and GCP, FedRAMP High, and NIST 800-53 high-impact controls — a design target it's engineered to satisfy, not an authorization it holds.