Strands Agents

# Strands Agents > Strands Agents is a simple yet powerful SDK that takes a model-driven approach to building and running AI agents. From simple conversational assistants to complex autonomous workflows, from local development to production deployment, Strands Agents scales with your needs. ## Docs - Get Started - [overview](https://strandsagents.com/docs/user-guide/quickstart/overview/index.md): Documentation for the Strands Agents SDK. Learn to build, deploy, and scale AI agents with tools, multi-agent patterns, and any model provider. - [Quickstart: Python](https://strandsagents.com/docs/user-guide/quickstart/python/index.md): Get an AI agent running in Python in under 5 minutes. Install the SDK, configure a model provider, and build your first agent. - [Quickstart: TypeScript](https://strandsagents.com/docs/user-guide/quickstart/typescript/index.md): Get an AI agent running in TypeScript in under 5 minutes. Install the SDK, configure a model provider, and build your first agent. - [build-with-ai](https://strandsagents.com/docs/user-guide/build-with-ai/index.md): Use AI coding assistants to build Strands Agents projects with up-to-date documentation context. - Build - [Adding Tools](https://strandsagents.com/docs/user-guide/concepts/tools/index.md): Give AI agents the ability to take real-world actions. Strands supports built-in tools, custom Python functions, MCP servers, and community packages. - [Creating Custom Tools](https://strandsagents.com/docs/user-guide/concepts/tools/custom-tools/index.md): Turn any Python function into an AI agent tool. Define inputs with docstrings, and the agent calls your function autonomously. - [Using MCP Tools](https://strandsagents.com/docs/user-guide/concepts/tools/mcp-tools/index.md): Connect AI agents to any MCP server for tool access. Strands agents integrate and invoke MCP tools without custom integration code. - [Multi-Agent Systems](https://strandsagents.com/docs/user-guide/concepts/multi-agent/multi-agent-patterns/index.md): Coordinate multiple AI agents with handoffs, swarms, graphs, and workflow patterns. Built-in primitives for complex multi-agent orchestration. - [Structured Output](https://strandsagents.com/docs/user-guide/concepts/agents/structured-output/index.md): Get typed, validated responses from AI agents. Define Pydantic models and Strands returns structured data instead of raw text. - [Streaming Responses](https://strandsagents.com/docs/user-guide/concepts/streaming/index.md): Stream AI agent responses in real time. Handle partial results, tool invocations, and token-by-token output as the agent works. - [Voice & Realtime](https://strandsagents.com/docs/user-guide/concepts/bidirectional-streaming/quickstart/index.md): Build voice-enabled AI agents with real-time audio streaming. Works with Amazon Nova Sonic, Gemini Live, and OpenAI Realtime. - Concepts - [agent-loop](https://strandsagents.com/docs/user-guide/concepts/agents/agent-loop/index.md): How Strands agents reason, plan, and act. Understand the model-driven loop that orchestrates tool use, reflection, and goal completion. - [state](https://strandsagents.com/docs/user-guide/concepts/agents/state/index.md): Maintain state in Strands agents across turns and sessions: conversation history, key-value agent state, and per-request invocation state. - [storage](https://strandsagents.com/docs/user-guide/concepts/storage/index.md): Choose a storage backend and pass it to plugins that need persistence: session managers, context offloaders, and memory. - [session-management](https://strandsagents.com/docs/user-guide/concepts/agents/session-management/index.md): Persist agent conversations across sessions. Save and restore chat history, tool state, and context for long-running AI workflows. - [snapshots](https://strandsagents.com/docs/user-guide/concepts/agents/snapshots/index.md): Capture and restore agent state at any point in time. Use snapshots for checkpointing, undo/redo, branching co […gekürzt]