Skip to main content

O2AGENT

O2AGENT is a Go-native platform for the full lifecycle management of AI agents — build, run, govern, observe, and monetize agents at enterprise scale.

Think of it as what API Management did for APIs, applied to autonomous and semi-autonomous AI agents: a single platform that takes an agent from "an idea a developer is prototyping" to "a governed, observable, monetizable capability running safely in production," without every team reinventing registries, gateways, sandboxes, and audit trails from scratch.

Why O2AGENT​

Teams building with AI agents today end up solving the same handful of hard problems over and over: how do you register and version an agent's configuration, safely execute the code it writes, enforce who (or what) is allowed to invoke it, trace what it actually did, and eventually let other teams — or paying customers — discover and use it? O2AGENT answers all of that with one coherent platform instead of a pile of bespoke scripts.

What O2AGENT gives you​

Build & Run​

  • Agent Registry & Catalog — register agents, version their configuration, and describe their capabilities (tools, models, permissions, cost profile) in one place.
  • Agent Runtime & Orchestration — a runtime engine that handles LLM calls, tool/function calling, memory, and multi-agent workflows (sequential, hierarchical, swarm), including durable long-running execution and sandboxed code execution for agents that write and run code as a tool.

Govern & Secure​

  • Gateway & Control Plane — authentication and authorization, rate limiting, policy enforcement, model-provider routing, and protocol translation (MCP, A2A, OpenAI-style tool calling) in front of every agent invocation.
  • Identity, Multi-Tenancy & Access Control — human users, RBAC roles/scopes, first-class AI agent identities, and human→agent delegation, so every action — human or agent — is properly authenticated and scoped.

Observe & Improve​

  • Observability & Evaluation — distributed tracing, token/cost metering, agent evaluations, guardrails, and audit logs, so you always know what your agents did and what it cost.

Extend & Build Faster​

  • Developer Experience — SDKs, a CLI with both scriptable command mode and an interactive TUI, a local test harness, and agent templates that get you from zero to a running agent quickly.
  • Marketplace & Ecosystem — publish, share, and monetize agents and connectors through a plugin system built for extensibility.

Deploy Anywhere​

  • Deployment & Operations — a Kubernetes-native self-hosted install and a multi-region hosted SaaS control plane, from the same codebase, so you choose where your agents run without changing how you build them.

Where to go next​

  • Getting Started — install the CLI, stand up the platform, and run your first agent.
  • Concepts — the core building blocks: agents, versions, tools, and identity.
  • CLI Reference — every o2agentctl command.
  • Architecture — how a request flows through the platform.