OpenAI Agents SDK

Plug MCP servers into this product so it can use their tools G

Agent access

Drive the product through a documented public API G

Agent access

Delegate tasks to a built-in AI assistant inside the product G

Agentic features

Connect an agent via an official MCP server G

Agent access

Build against official SDKs G

Agent access

Point an agent at llms.txt or agent-oriented docs G

Agent access

Operate the product with natural-language commands G

Agentic features

Run the product headlessly / in CI for automation G

Agent access

Download a machine-readable API spec (OpenAPI or equivalent) G

Api quality

Explore an interactive API reference with runnable examples G

Api quality

Rely on versioned APIs with a documented deprecation policy G

Api quality

Set up automations that run autonomously in the background G

Agentic features

Use an official CLI G

Agent access

Get AI-generated insights and suggestions from my data inside the product G

Agentic features

Issue scoped/least-privilege API credentials for an agent G

Agent access

Subscribe to events via webhooks G

Agent access

Test against a sandbox environment without touching production data G

Api quality

Define an agent with typed custom tools in a few lines of code C

Agent authoring

Orchestrate multiple agents — handoffs, subagents, or crews — inside one workflow C

Multi agent

Stream tokens and intermediate agent events (tool calls, steps) to my UI in real time C

Streaming

Trace every LLM call and tool invocation of an agent run in an observability UI C

Tracing

Attach input/output guardrails that validate, transform, or block unsafe content C

Guardrails

Pause an agent mid-run for human input or approval and resume with the human's decision C

Approval flows

Self-host the core product G

Swap the underlying LLM provider or model without rewriting my agent C

Portability

Checkpoint agent state so a run can resume exactly where it left off after a crash or restart C

Durable state

Get schema-validated structured output from an agent, with automatic retries when validation fails C

Structured output

Define rules that trigger actions automatically on events G

Export all of my data in open formats and leave G

Prevent my data from being used to train AI models G

Require human approval before specific sensitive tool calls execute C

Approval flows

Restrict what an agent may do with fine-grained tool permissions and sandboxed execution C

Guardrails

Rely on strict typing and schema validation so a coding agent catches its own mistakes at build time C

Ai buildability

Run my agents entirely on my own infrastructure with no dependence on the vendor's platform C

Deployment

Compose agents into an explicit graph or workflow with branching, loops, and parallel steps C

Workflow control

Give agents long-term memory that persists across sessions and threads C

Memory

Read the product's source under an open license G

Run the framework's example agents headlessly from a terminal so an agent can verify what it just built C

Ai buildability

Run long-lived agents durably across process restarts and deploys, natively or via durable-execution integrations C

Durable state

Deploy an agent to a managed runtime and call it as an API endpoint C

Deployment

Opt out of telemetry and usage tracking G

Trim, summarize, or filter conversation history to keep an agent inside its context window C

Memory

Choose where my data is stored (region/residency) G

Control data retention and deletion G

Do everything through the API that I can do in the UI G

Have a coding agent scaffold a new agent project from an official CLI or template in one command C

Ai buildability

Perform bulk operations across many items at once G

Schedule recurring jobs or workflows G

Score agent quality with built-in evals and run them as part of CI C

Evals

Unit-test agents with mocked models and tools C

Testing

Version, review, and roll back my automations G