# OpenAgentCore > The OpenAI Agents API, open source and self-hosted. Run Codex, Claude Code or MiniMax Code on your own machines with your own models, through the official OpenAI SDK. ## Docs › Get started - [Overview](/docs/getting-started/index.md): OpenAgentCore runs AI agents on your own infrastructure behind the OpenAI Agents API. - [Install Core and Web](/docs/getting-started/install.md): This guide takes an operator from an empty host to a working installation: Core, the Web console and PostgreSQL running under Docker, a sandbox backend, a default model and a fi… - [Run your first Session](/docs/getting-started/quickstart.md): Run an agent through Core with the official OpenAI Python SDK: connect with a Project API key, start a Session that writes a file, and read the result. ## Docs › Operate - [Installation options](/docs/getting-started/install-options.md): Reference for the installer's flags, and how to install differently from the default: behind an existing reverse proxy, with Docker Compose on a hosting platform, or on a host w… - [Configure the public address](/docs/getting-started/public-address.md): Set the one origin that browsers, nodes, sandboxes and applications use to reach Core, and put a reverse proxy with HTTPS in front of it. - [Web console](/docs/web/index.md): Web is the administrator console of an OpenAgentCore installation: operators sign in with the Core key to watch health and usage, inspect what applications created, and manage P… - [Console server](/docs/web/console-server.md): The console server (oac-web) serves Web, signs the administrator in with the Core key and forwards the browser's /core/v1 requests to Core. - [Choose a sandbox backend](/docs/getting-started/sandbox-backends.md): Compare microsandbox and E2B, then select one in Web. - [Set up workspace storage](/docs/getting-started/workspace-storage.md): Keep Session workspaces on an NFS filesystem outside the sandbox nodes, with or without per-workspace quotas. - [Add and manage nodes](/docs/getting-started/nodes.md): A node is a Linux host that runs sandboxes for Core-hosted Sessions. - [Operations](/docs/getting-started/operations.md): Day-to-day tasks for the operator of a Core installation: the oac command, service health, the Core key, backups and uninstalling. - [Troubleshooting](/docs/getting-started/troubleshooting.md): Find the symptom, its cause and the fix for installation, access, nodes and Sessions. - [Configuration reference](/docs/configuration.md): Every setting of a Core installation, where it lives and how to change it. ## Docs › Concepts - [Concepts](/docs/concepts.md): A Project is the execution tenant in OpenAgentCore. - [Architecture](/docs/architecture.md): OpenAgentCore is an Agent execution platform that brings Harnesses, models, tools, Sessions and execution environments together through replaceable protocol implementations. - [Performance](/docs/performance.md): These measurements cover the sandbox lifecycle on microsandbox: starting a sandbox, preparing its environment, suspending it when idle and resuming it. ## API › Build on the API - [API namespaces and credentials](/docs/api/index.md): Core serves three API namespaces, each for one kind of caller with its own credential: applications on /v1, administrators on /core/v1, and machines on /api/v1. - [Agents API guide](/docs/api/public-agent-api.md): Core serves the OpenAI Agents API at /v1. - [Run self-hosted Sessions](/docs/getting-started/self-hosted.md): A selfhosted Session runs on a machine your application owns, such as a workstation, VM or container, instead of a sandbox Core manages. - [Examples](/docs/examples.md): Complete applications built on the Agents API guide, each running against a real Core installation with a Project API key. ## API › Agents API reference - [OpenAI compatibility](/contracts/agents-api/index.md): Core implements the OpenAI Agents API pinned below. - [API behavior](/contracts/agents-api/wire-semantics.md): The pinned OpenAI Python SDK (upstream.json) defines the /v1 routes, fields and types. - [Harness capabilities](/contracts/agents-api/harness-capabilities.md): This page lists what each Harness supports on each placement. - [Sessions, events and history](/contracts/agents-api/sessions-events.md): This contract covers what happens inside a Session: sending input, the live event stream, and reading the durable history of Turns, Items and usage. - [Message content](/contracts/agents-api/message-content.md): User messages and function results share one content model: an ordered list of inputtext and inputimage parts. - [Execution tools](/contracts/agents-api/execution-tools.md): An Agent declares application functions, controls and MCP servers in tools, and an optional output schema in text.format. - [Subagents](/contracts/agents-api/subagents.md): With multiagent.enabled, a Harness may start native child agents. - [Environments and Templates](/contracts/agents-api/environments.md): An Environment is the execution resource of a Session: the machine, workspace and prepared capabilities that a Harness runs in. - [Environment files and Artifacts](/contracts/agents-api/environment-files.md): A Session's workspace holds live files that the agent and its tools change. - [Files and Skills](/contracts/agents-api/source-files.md): Files (/v1/files) and Skills (/v1/skills) are Project resources with their own lifecycle, independent of Sessions. - [Vaults and Credentials](/contracts/agents-api/vaults.md): A Vault is a Project-owned container of Credentials. ## API › Core API reference - [Core administration API](/contracts/agents-api/admin-api.md): The Core administration API (/core/v1) manages an installation: Projects and their API keys, reads and deletion of Project resources, executor credentials, deployment default mo… - [Core administration errors](/contracts/agents-api/core-errors.md): Errors on /core/v1 use this envelope. - [Sandbox deployment](/contracts/agents-api/sandbox-deployment.md): The sandbox deployment selects the Sandbox Provider and the per-sandbox resources for Core-managed openaihosted execution. - [Executor credentials](/contracts/agents-api/environment-executor-credentials.md): An executor credential lets oac-daemon enroll and connect for one selfhosted Environment. - [Session diagnostics](/contracts/agents-api/session-diagnostics.md): Administrators read why a Session or Turn failed through two Core-key routes that report failure categories and Item receipt timing. - [Core metrics](/contracts/agents-api/core-metrics.md): GET /core/v1/metrics?range=1h|6h|24h|7d reports Core's own health: its process, execution queue and slots, PostgreSQL and background jobs. - [Runtime telemetry API](/contracts/agents-api/runtime-observability-api.md): Core reports what hosted Runtimes and sandbox nodes consume through read-only administrator routes under /core/v1: current Runtime observations, the stored Runtime history of on… - [Runtime observability](/contracts/agents-api/runtime-observability.md): This is the contributor contract for how Core observes Runtimes and keeps their history. - [Machine connection API](/contracts/agents-api/machine-api.md): Machines call Core under /api/v1: sandbox nodes, Runtime daemons and the self-hosted installer. ## Extend › Protocols - [Core–Runtime protocol](/docs/runtime-protocol.md): The Core–Runtime protocol connects Core to a Runtime daemon and defines the meaning and order of the messages on the daemon connection. - [Runtime bootstrap](/docs/runtime-bootstrap.md): The Runtime bootstrap protocol connects a Sandbox Provider to the managed Runtime it starts. - [Harness protocol](/contracts/agents-api/harness-onboarding.md): A Harness is a native agent engine (Codex, Claude Code, MiniMax Code) that runs the model and tool loop; a Harness adapter translates the Runtime's Executor and Turn contract in… - [Built-in Harness registrations](/contracts/agents-api/harness-catalog.md): Each Harness is the one declaration.go in its internal/harnessconfig/ directory. - [Model execution](/contracts/agents-api/model-execution.md): Each Session runs one Harness with one model provider. - [Sandbox Provider protocol](/docs/sandbox-provider.md): The Sandbox Provider protocol connects Core to the adapter that supplies compute for Core-managed (openaihosted) Environments and runs the bounded bootstrap that starts the Runt… - [Sandbox node protocol](/contracts/agents-api/node-generation-protocol.md): The sandbox node protocol connects Core to a microsandbox host, which runs Provider operations against its local provider and reports readiness, host measurements and the deploy… - [Workspace filesystem protocol](/docs/workspace-provider.md): The workspace filesystem protocol connects Core and sandbox nodes to the adapter that owns durable Environment storage. ## Extend › Contribute - [Develop OpenAgentCore](/docs/development.md): Set up a checkout, build a component and validate your changes. - [Build and release OpenAgentCore](/docs/maintainers.md): This guide is for maintainers who build and publish OpenAgentCore. - [Console API usage](/docs/web/console-api-usage.md): This page lists the Core routes each console page calls and how the console bounds its reads.