Related Products
What else should we look at alongside Claude Agent SDK?
Alternatives
Other frameworks teams evaluate for running agents in their own code.
- OpenAI Agents SDKAgent Framework−17 MixedOpenAI's Python and TypeScript SDK for agents, handoffs, guardrails, and sessions, with tracing built in.
- LangGraphAgent Framework+25 PositiveLangChain's MIT-licensed runtime for stateful agents built as graphs, with checkpoints and interrupts across model providers.
- AI SDKAgent Framework+36 PositiveVercel's TypeScript toolkit for model calls, tool loops, and streaming UI across providers.
Complements
Where the SDK's model calls are served and where its traces go.
- Claude APIAI Model API+25 PositiveAnthropic's developer platform, which bills the SDK's tokens and offers Managed Agents and the tool runner for teams that skip the subprocess.
- Amazon BedrockAI Model API+9 MixedAWS's model platform, which serves Claude to the SDK when CLAUDE_CODE_USE_BEDROCK is set.
- LangfuseAI Observability+46 PositiveOpen-source LLM tracing that can receive the OpenTelemetry data the SDK's CLI process exports.
