# Technical Documentation: strands-agents/sdk-python > ℹ️ **Provenance:** Hybrid Fusion: `strands-agents/sdk-python` (README + 3 In-Tree Chapters) · [CodeWiki Reference](https://codewiki.google/github.com/strands-agents/sdk-python) · Recency: Active (< 180 days) ## 1. Project Overview & Quickstart (strands-agents/sdk-python)

Strands Agents

A model-driven approach to building AI agents in just a few lines of code.

Documentation ◆ Samples ◆ Tools ◆ MCP Server

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. This monorepo contains the Python SDK, TypeScript SDK, documentation site, and supporting packages: | Directory | Description | |-----------|-------------| | `strands-py/` | Python SDK: agent loop, model providers, tools ([PyPI](https://pypi.org/project/strands-agents/) · [releases](https://github.com/strands-agents/harness-sdk/releases?q=python%2F&expanded=false)) | | `strands-ts/` | TypeScript SDK: agent loop, model providers, tools ([npm](https://www.npmjs.com/package/@strands-agents/sdk) · [releases](https://github.com/strands-agents/harness-sdk/releases?q=typescript%2F&expanded=false)) | | `site/` | Source for the [strandsagents.com](https://strandsagents.com) documentation site (Astro/Starlight) | | `team/` | Governance and cross-SDK process docs (tenets, decisions, PR & compatibility guidelines, and `designs/` proposals) | ## Why Strands Build an agent harness. Control it end-to-end. - **Build your way.** Any model, any cloud. Context management, execution limits, and observability built in before you write a line of config. Swap backends when you scale; your code stays the same. - **Model agnostic.** First-class support for Amazon Bedrock, Anthropic, OpenAI, and Gemini, plus [many more providers](https://strandsagents.com/docs/user-guide/concepts/model-providers/) and custom ones. - **Stay in control.** The agent loop traces every decision by default. Hooks let you intercept any step to log it, validate it, or redirect it. - **Deliver outcomes that work.** Guardrails catch mistakes before they run. Steering handlers let agents correct themselves instead of failing silently. MCP, streaming, multi-agent patterns, and structured output are all built in. ## Quick Start Both SDKs default to the Amazon Bedrock model provider, so you'll need AWS credentials configured and model access enabled for Claude Sonnet. The [Quickstart Guide](https://strandsagents.com/docs/user-guide/quickstart/overview/) covers configuring other providers (Anthropic, OpenAI, Gemini, Ollama, and more). ### Python Requires Python 3.10+: ```bash pip install strands-agents strands-agents-tools ``` ```python from strands import Agent from strands_tools import calculator agent = Agent(tools=[calculator]) agent("What is the square root of 1764") ``` The [Python SDK README](strands-py/) covers tools, model providers, MCP, and bidirectional streaming. ### TypeScript Requires Node.js 20+: ```bash npm install @strands-agents/sdk ``` ```typescript import { Agent } from '@strands-agents/sdk' const agent = new Agent() const result = await agent.invoke('What is the square root of 1764?') console.log(result) ``` More in the [TypeScript SDK README](strands-ts/), including Zod-typed tools, structured output, and multi-agent patterns. ## Documentation For detailed guidance & examples, explore our documentation: - [User Guide](https://strandsagents.com/) - [Quick Start Guide](https://strandsagents.com/docs/user-guide/quickstart/overview/) - [Agent Loop](https://strandsagents.com/docs/user-guide/concepts/agents/agent-loop/) - [Examples](https://strandsagents.com/docs/examples/) - API Reference: [Python](https://strandsagents.com/docs/api/python/strands.agent.agent/) · [TypeScript](https://strandsagents.com/docs/api/typescript/) - [Production & Deployment Guide](https://strandsagents.com/docs/user-guide/deploy/operating-agents-in-production/) The docs themselves live in this monorepo under [`site/`](site/), and doc PRs are welcome alongside code changes. ## Development Git operations (commits, branches, PRs) are done from the repo root. Each package has its own toolchain: **Python SDK** (`strands-py/`): ```bash cd strands-py pip install hatch hatch test # run unit tests hatch fmt # format & lint ``` **TypeScript SDK** (`strands-ts/`): ```bash npm ci # install from repo root npm run build # build npm test # run unit tests ``` **Documentation site** (`site/`): ```bash cd site npm install npm run dev # local dev server at http://localhost:4321/ ``` ## Contributing ❤️ We welcome contributions! See our [Contributing Guide](CONTRIBUTING.md) for details on: - Reporting bugs & features - Development setup - Contributing via Pull Requests - Code of Conduct - Reporting of security issues ## Stay in touch with the team Come meet the Strands team and other users on [**Discord**](https://discord.com/invite/strands) ## License This project is licensed under the Apache License 2.0 - see the [LICENSE.APACHE](LICENSE.APACHE) file for details. ## Security See [CONTRIBUTING](CONTRIBUTING.md#security-issue-notifications) for more information. ## 2. In-Tree Documentation Chapters (strands-agents/sdk-python) ### CONTRIBUTING # Contributing Guidelines Thank you for your interest in contributing to our project. Whether it's a bug report, new feature, correction, or additional documentation, we greatly value feedback and contributions from our community. Please read through this document before submitting any issues or pull requests to ensure we have all the necessary information to effectively respond to your bug report or contribution. ## Reporting Bugs/Feature Requests We welcome you to use the [Bug Reports](../../issues/new?template=bug_report.yml) file to report bugs or [Feature Requests](../../issues/new?template=feature_request.yml) to suggest features. For a list of known bugs and feature requests: - Check [Bug Reports](../../issues?q=is%3Aissue%20state%3Aopen%20label%3Abug) for currently tracked issues - See [Feature Requests](../../issues?q=is%3Aissue%20state%3Aopen%20label%3Aenhancement) for requested enhancements When filing an issue, please check for already tracked items Please try to include as much information as you can. Details like these are incredibly useful: * A reproducible test case or series of steps * The version of our code being used (commit ID) * Any modifications you've made relevant to the bug * Anything unusual about your environment or deployment ## Finding contributions to work on Looking at the existing issues is a great way to find something to contribute to. We label issues that are well-defined and ready for community contributions with the "ready for contribution" label. Check our [Ready for Contribution](../../issues?q=is%3Aissue%20state%3Aopen%20label%3A%22ready%20for%20contribution%22) issues for items you can work on. Before starting work on any issue: 1. Check if someone is already assigned or working on it 2. Comment on the issue to express your interest and ask any clarifying questions 3. Wait for maintainer confirmation before beginning significant work ## Development Tenets Our team follows these core principles when designing and implementing features. These tenets help us make consistent decisions, resolve trade-offs, and maintain the quality and coherence of the SDK. When contributing, please consider how your changes align with these principles: 1. **Simple at any scale:** We believe that simple things should be simple. The same clean abstractions that power a weekend prototype should scale effortlessly to production workloads. We reject the notion that enterprise-grade means enterprise-complicated - Strands remains approachable whether it's your first agent or your millionth. 2. **Extensible by design:** We allow for as much configuration as possible, from hooks to model providers, session managers, tools, etc. We meet customers where they are with flexible extension points that are simple to integrate with. 3. **Composability:** Primitives are building blocks with each other. Each feature of Strands is developed with all other features in mind, they are consistent and complement one another. 4. **The obvious path is the happy path:** Through intuitive naming, helpful error messages, and thoughtful API design, we guide developers toward correct patterns and away from common pitfalls. 5. **We are accessible to humans and agents:** Strands is designed for both humans and AI to understand equally well. We don’t take shortcuts on curated DX for humans and we go the extra mile to make sure coding assistants can help you use those interfaces the right way. 6. **Embrace common standards:** We respect what came before, and do not want to reinvent something that is already widely adopted or done better. When proposing solutions or reviewing code, we reference these principles to guide our decisions. If two approaches seem equally valid, we choose the one that best aligns with our tenets. The `team/` folder holds the rest of our shared context: the full [tenets](./team/TENETS.md), [decision records](./team/DECISIONS.md), the [API bar-raising](./team/API_BAR_RAISING.md) and [feature lifecycle](./team/FEATURE_LIFECYCLE.md) processes, and our [design proposals](./team/designs/). For a significant feature or a breaking change, start with a design proposal in [`team/designs/`](./team/designs/) — see its [README](./team/designs/README.md) for the template and process. ## Development Environment This is a monorepo containing the Python SDK, TypeScript SDK, MCP server, and documentation site. Each has its own toolchain: | Area | Directory | Toolchain | |------|-----------|-----------| | Python SDK | `strands-py/` | hatch | | TypeScript SDK | `strands-ts/` | npm workspace | | MCP server | `strands-mcp/` | hatch | | Docs site | `site/` | Astro (npm) | ### Python SDK The Python SDK uses [hatchling](https://hatch.pypa.io/latest/build/#hatchling) as the build backend and [hatch](https://hatch.pypa.io/latest/) for development workflow management. All `hatch` commands should be run from the `strands-py/` subdirectory (where `pyproject.toml` lives): ```bash cd strands-py ``` ### Setting Up Your Development Environment 1. Entering virtual environment using `hatch` (recommended), then launch your IDE in the new shell. ```bash hatch shell ``` 2. Set up pre-commit hooks: ```bash pre-commit install -t pre-commit -t commit-msg ``` This will automatically run formatters and conventional commit checks on your code before each commit. 3. Run code formatters manually: ```bash hatch fmt --formatter ``` 4. Run linters: ```bash hatch fmt --linter ``` 5. Run unit tests: ```bash hatch test ``` Or run them with coverage: ```bash hatch test -c ``` 6. Run integration tests: ```bash hatch run test-integ ``` ### Pre-commit Hooks We use [pre-commit](https://pre-commit.com/) to automatically run quality checks before each commit. The hook will run `hatch run format`, `hatch run lint`, `hatch run test`, and `hatch run cz check` when you make a commit, ensuring code consistency. The pre-commit hook is installed with: ```bash pre-commit install ``` You can also run the hooks manually on all files: ```bash pre-commit run --all-files ``` ### Code Formatting and Style Guidelines We use the following tools to ensure code quality: 1. **ruff** - For formatting and linting 2. **mypy** - For static type checking These tools are configured in the [pyproject.toml](./strands-py/pyproject.toml) file. Please ensure your code passes all linting and type checks before submitting a pull request: ```bash # Run all checks hatch fmt --formatter hatch fmt --linter ``` If you're using an IDE like VS Code or PyCharm, consider configuring it to use these tools automatically. For additional details on styling, please see our dedicated [Style Guide](./strands-py/docs/STYLE_GUIDE.md). ### TypeScript SDK The TypeScript SDK uses an npm workspace rooted at the repository root. ```bash npm ci # install dependencies (from repo root) npm run build # build npm test # run unit tests npm run lint # lint npm run type-check # type checking ``` #### Running Selective Integration Tests Locally From the repository root, run only the integration tests relevant to your changes (computed relative to `main`, including uncommitted edits): ```bash npm run test:integ:selective ``` This uses Vitest's module graph to run only the `integ-node` and `integ-browser` specs that depend on the source files you changed. If you alter a structural file (`package.json`, `package-lock.json`, a `strands-ts` `tsconfig`, `vitest.config.ts`, a shared integration fixture under `test/integ/__fixtures__/`, or a TypeScript CI workflow), the full integration suite runs automatically. ### Documentation Site The documentation site uses Astro with the Starlight theme. ```bash cd site npm install npm run dev # local dev server at http://localhost:4321/ npm run build # production build npm run typecheck # type checking npm run typecheck:snippets # type check code examples ``` For docs contribution guidelines, see [site/CONTRIBUTING.md](./site/CONTRIBUTING.md). ### PR Size and Complexity Every PR is labeled with a `size/*` and a `complexity/*` label. Both are informational — they help reviewers budget attention and neither blocks a merge. You can see the same numbers before you push. From the repository root: ```bash npm run complexity:setup # once, to install the analyzers npm run complexity # report labels for your branch vs origin/main ``` Python contributors can use hatch instead, from `strands-py/`: ```bash cd strands-py hatch run complexity ``` **`size/*`** counts changed lines in source and prose, and **excludes tests, lockfiles, and snapshots**. Thorough tests should never push a PR into a bigger bucket, so write as many as the change deserves. **`complexity/*`** reports the [cognitive complexity](https://www.sonarsource.com/docs/CognitiveComplexity.pdf) of the most complex function your diff touches — roughly, how hard the control flow is to hold in your head. It scores only the functions you actually changed, so an existing hotspot elsewhere in a file you edited will not count against you. `complexity/high` (above 25) is a hint that a function may be worth splitting, not a rule; sometimes a complex function is the honest solution. A docs-only or test-only PR touches no SDK source and gets no complexity label. The labels exist because review attention is the scarcest resource this project has: they let maintainers triage which PRs need an unhurried senior read, and they give you the same signal locally before a reviewer sees it. The reasoning, the scoring model, and the practices that keep code under the thresholds live in [team/COMPLEXITY.md](./team/COMPLEXITY.md). ## Using AI Tools We love AI. We build with coding agents every day, and you're welcome to use them too — they're a great way to move fast and explore a codebase. That said, **you are the author of your pull request, not your agent.** Before you open a PR, make sure you understand the code well enough to explain why it works, defend the design choices, and maintain it if asked. If you couldn't walk a reviewer through it line by line, it's not ready yet. A few things that help us help you: - **Keep changes small and incremental.** A focused PR that does one thing is far easier for us to understand, guide, and merge than a large one that touches many areas. When in doubt, split it up. - **Open an issue first for anything significant**, so we can align on the approach before you (or your agent) invest the time. - **Review every line your agent generates.** Delete what you don't need, simplify what's over-engineered, and make sure tests actually exercise the behavior — not just pass. Trim comments that narrate the agent's reasoning: a comment should state only what a reader cannot infer from the code. High-quality PRs get reviewed faster and are far more likely to be accepted. Taking the time to understand and trim your changes is the single best thing you can do to get them merged. ## Contributing via Pull Requests Contributions via pull requests are much appreciated. Before sending us a pull request, please ensure that: 1. You are working against the latest source on the *main* branch. 2. You check existing open, and recently merged, pull requests to make sure someone else hasn't addressed the problem already. 3. You open an issue to discuss any significant work - we would hate for your time to be wasted. For guidance on writing effective PR descriptions, see our [PR Description Guidelines](./team/PR.md). To send us a pull request, please: 1. Create a branch. 2. Modify the source; please focus on the specific change you are contributing. If you also reformat all the code, it will be hard for us to focus on your change. 3. Format your code using `hatch fmt --formatter`. 4. Run linting checks with `hatch fmt --linter`. 5. Ensure local tests pass with `hatch test` and `hatch run test-integ`. 6. Commit to your branch using clear commit messages following the [Conventional Commits](https://www.conventionalcommits.org) specification. 7. Send us a pull request, answering any default questions in the pull request interface. 8. Pay attention to any automated CI failures reported in the pull request, and stay involved in the conversation. ## Code of Conduct This project has adopted the [Amazon Open Source Code of Conduct](https://aws.github.io/code-of-conduct). For more information see the [Code of Conduct FAQ](https://aws.github.io/code-of-conduct-faq) or contact opensource-codeofconduct@amazon.com with any additional questions or comments. ## Security issue notifications If you discover a potential security issue in this project we ask that you notify AWS/Amazon Security via our [vulnerability reporting page](http://aws.amazon.com/security/vulnerability-reporting/). Please do **not** create a public github issue. ## Licensing See the [LICENSE.APACHE](./LICENSE.APACHE) and [LICENSE.MIT](./LICENSE.MIT) files for our project's licensing. We will ask you to confirm the licensing of your contribution. --- ### README

Strands Agents

A model-driven approach to building AI agents in just a few lines of code.

Documentation ◆ Samples ◆ Tools ◆ MCP Server