Skip to content
Agents API
Rebyte Docs
Agents API
OverviewQuickstartArchitectureConfiguring Agents
Sessions
Run and continue sessionsEvents and itemsManage sessionsWebhooks
Environments and sandboxes
Rebyte SandboxesSandbox lifecycleSandbox securityFiles and artifacts
Tools and integrations
Web searchFunctionsMCP connectionsPluginsVaults
Multi-agentObservability and usageTracing

Not yet supported

Agents API

Overview

Build applications with a managed agent runtime, durable sessions, and tools.

pnpm add @rebyteai/agent-sdk

Use Node.js 22+ and set REBYTE_API_KEY. The SDK connects to Rebyte by default; no repository clone or base URL configuration is required.

Our two application repositories serve different purposes:

SDKrebyte-agent-toolkitThe Agent SDK and AppKit: API client, React hooks, chat UI, server adapter and CLI. Install the packages from npm.Democommerce-agent-starterA complete commerce demo with catalog, cart and client functions. Clone it to run or customize the storefront.
A familiar SDK
Compatible with the OpenAI Agents API client. Our maintained SDK fork connects to Rebyte by default—just add your API key. View supported features
Trace execution
Follow model calls, tool activity and token usage with execution traces tied to each session and turn. Explore observability
Control access
Isolate session files, restrict sandbox network access and keep service MCP credentials outside the sandbox in encrypted Vaults. Runtime security
Run your first session SDK and example app
On this pageCore conceptsFrom input to resultCurrent availabilityRebyte Agent SDKSDK and example applicationRebyte AppKit SDKCommerce Agent

Core concepts

ConceptPurpose
AgentReusable configuration: model, instructions, tools, and generation settings.
EnvironmentOptional compute and filesystem used by a session.
SessionPersistent execution and conversation state, with a resolved copy of the agent configuration.
Events and itemsEvents communicate inputs and progress; items retain messages and execution output.

A Turn tracks one unit of work inside a session. An Artifact is an immutable deliverable produced by a completed turn. A saved Agent can be reused across Sessions; each managed Session owns its own environment.

From input to result

  1. Create a Session with an inline Agent definition or a saved Agent ID.
  2. Provide input at creation or submit it through the Session events endpoint.
  3. Receive text, tool activity, and lifecycle events as the runtime works.
  4. Read the resulting Items, Turn status, and Artifacts.
  5. Continue the Session with another task, or queue another task behind its active Turn.

Your application handles its UI and function tools. Rebyte manages execution and the Session's hosted environment. See Architecture.

Your application starts sessions and receives events and output from the Agents API. Rebyte runs the managed agent harness, which sends tool calls to the Sandbox and receives tool results. Rebyte provisions and manages both the runtime and the Sandbox.

Current availability

The production endpoint is https://api.rebyte.ai/v1. Use http://127.0.0.1:34567/v1 only with a running local Relay and a key for its local organization.

CapabilityCurrent implementation
Agents, Sessions, Turns, ItemsImplemented
Live event streaming and input queuingImplemented; historical events are not replayed; active-Turn input queues
Managed environments, inline files, ArtifactsImplemented
Functions, HTTP and stdio MCP, VaultsImplemented with documented limits
Web searchImplemented; live mode only, no cached search or location targeting
Inline ZIP skills, plus Rebyte's GitHub-sourced skillsImplemented in managed environments
Webhooks, PluginsNot yet supported
Execution tracingImplemented through Langfuse; view Agent, Session and Turn traces in Platform
Multi-agentNot yet supported
Automatic context compactionNot yet supported

The navigation follows OpenAI's guide. A hollow dot marks an unsupported capability. Compatibility is defined by the implemented fields and behavior, not by the presence of a page in this guide.

Rebyte Agent SDK

Rebyte maintains a source fork of the OpenAI Agents API TypeScript client. Use the same client.beta.agents methods, streaming events and resource model, with Rebyte credentials and a Rebyte endpoint configured by default.

import Rebyte, { rebyteSandbox } from '@rebyteai/agent-sdk';

const client = new Rebyte(); // reads REBYTE_API_KEY
const session = await client.beta.agents.sessions.create({
  agent: { model: 'gpt-5.6-luna', instructions: 'Help the user.' },
  environment: rebyteSandbox(),
  input: 'Hello!',
});

Install the published SDK directly in your application; follow the Quickstart. No repository clone is required. rebyteSandbox() configures Rebyte-managed compute. Omit environment when you only need the Agent Loop, service MCP or client functions. Existing OpenAI API clients remain compatible with supported endpoints. This fork is based on openai@7.15.0; it is distinct from @openai/agents, which runs an Agent Loop in your application. Python integrations can continue using openai==3.13.0 with Rebyte's base URL; a Python fork is not included in this release.

The SDK sets OpenAI-Beta: agents=v1 automatically. Raw HTTP clients must supply that header. Current availability describes server support.

SDK and example application

RepositoryWhat it providesHow to use it
rebyte-agent-toolkitReusable Agent SDK, AppKit and CLI, plus Node and Cloudflare examples.Install packages from npm. Clone only to run the complete examples or contribute source changes.
commerce-agent-starterA complete commerce demo with catalog, cart, Skills and client functions.Clone it to run or customize the example storefront.

Rebyte AppKit SDK

Rebyte AppKit is Rebyte's application SDK, published in the Rebyte Agent Toolkit repository. It provides @rebyteai/agent-react for React session state, @rebyteai/agent-ui for a ready-made chat interface, @rebyteai/agent-server for the server adapter, and a configuration CLI.

Start with the Node application or Cloudflare application. The Rebyte SDK recipes cover Agent creation, independent Sessions, streaming, files and client functions. Install only the packages your application needs:

npm packagePurpose
@rebyteai/agent-sdkTypeScript API client for Agents, Sessions, streaming and Artifacts.
@rebyteai/agent-reactReact hooks and session state.
@rebyteai/agent-uiReady-made React chat interface and styles.
@rebyteai/agent-serverServer adapter that keeps your Rebyte API key out of the browser.
@rebyteai/cliCreate, update, validate and export Agent configuration from the terminal.
pnpm add @rebyteai/agent-sdk

All five packages are published on npm. The same repository contains their source; you do not need to clone it to use the SDK or AppKit.

Commerce Agent

Commerce Agent is a complete example storefront. Its Python host implements catalog, cart and presentation functions, with a separate API Session for each conversation and per-Session Skills. Follow the Rebyte integration guide for setup and the execution flow.

AppKit uses the Rebyte Agent SDK fork. Commerce currently uses the official Python client against Rebyte’s compatible Agents API.

API Agents appear in Platform. They are independent of product UI Agent Profiles and Workspaces. A Session never inherits another Session's Sandbox or Artifacts because it uses the same saved Agent.

To list your organization's saved API Agents, use GET /v1/agents or client.beta.agents.list(). See Agent listing and pagination for authentication, a cURL request, and an SDK example that retrieves every page.

Protocol referenceOpenAI Agents API ↗