# Strands Agents review

> Strands Agents is an open-source SDK for building and running AI agents in Python and TypeScript, with any model and any cloud.

Devtoolwatch Board, the week of Oct 25–31, 2026. Canonical page: https://index.devtoolwatch.com/tools/strands/

## TL;DR

Strands Agents is an open-source SDK for building and running AI agents in Python and TypeScript. It is for developers who would otherwise write their own agent loop. It runs in your process with no hosted control plane. It includes lifecycle controls, tools, MCP, multi-agent patterns, memory, sessions, guardrails, tracing, and evals in one SDK.

## Strands Agents at a glance

- **Category:** [Agent frameworks](https://index.devtoolwatch.com/categories/agent-frameworks/)
- **Maintained by:** strands-agents (organization, [@strands-agents](https://github.com/strands-agents) on GitHub)
- **License:** Apache 2.0
- **Language:** Python, TypeScript, MDX
- **Repository:** [github.com/strands-agents/harness-sdk](https://github.com/strands-agents/harness-sdk)
- **Website:** [strandsagents.com](http://strandsagents.com/)
- **PyPI package:** [strands-agents](https://pypi.org/project/strands-agents/) 1.62.0
- **Latest release:** [harness-python/v0.3.2](https://github.com/strands-agents/harness-sdk/releases/tag/harness-python/v0.3.2) on GitHub, Oct 22, 2026
- **First published:** May 14, 2025

## What Strands Agents does

- Strands Agents provides lifecycle controls for the agent loop: turn limits, token budgets, cancellation, and stop reasons.
- Strands Agents supports tools, structured output, MCP, streaming, and multi-agent patterns as built-in features.
- Strands Agents works with Amazon Bedrock, Anthropic, OpenAI, Gemini, Ollama, many other providers, and custom ones.
- Strands Agents traces every decision in the agent loop by default. Hooks let you intercept any step to log, validate, or redirect it.
- Strands harness gives you a fully assembled agent with defaults for model, tools, memory, sessions, and context management from one call.
- Strands Agents includes guardrails and steering handlers so agents can correct themselves instead of failing silently.

### How it works

Strands Agents runs inside your own Python or TypeScript process, with no hosted control plane. You can start with Strands harness via create_harness() or createHarness(), or use the SDK to own the agent loop and wire tools, model providers, and memory yourself. It connects to model providers such as Amazon Bedrock, Anthropic, OpenAI, and Gemini, and to MCP servers. A strands CLI lets you chat with a harness agent from the terminal.

## How to install Strands Agents

**Strands harness with pip**

```
pip install strands-harness
```

**Python SDK with pip**

```
pip install strands-agents strands-agents-tools
```

**TypeScript SDK with npm**

```
npm install @strands-agents/sdk
```

### Requirements

- Python 3.10 or later for the Python SDK.
- Node.js 22 or later in an ESM project ("type": "module") for the TypeScript SDK.
- Credentials for a model provider. The TypeScript SDK defaults to Amazon Bedrock, which needs AWS credentials.

### First steps

1. Install Strands harness and call create_harness() in Python, or createHarness() in TypeScript, to get an assembled agent.
2. Invoke the agent with a prompt, such as asking it to find the slowest test in your repo.
3. Configure AWS credentials or pass another model, because the TypeScript SDK's new Agent() uses Amazon Bedrock by default.
4. Move to the SDK when you want to own the agent loop, tools, model providers, and memory yourself.

## Who Strands Agents is for

- Developers who would otherwise hand-roll their own agent loop in Python or TypeScript.
- Teams that want to switch model providers or clouds without rewriting agent code.
- Developers deploying agents to AWS services such as Bedrock AgentCore, Lambda, Fargate, or EKS.
- Teams that need tracing, guardrails, and evals for agents in production.

### Limitations

- Strands Agents' TypeScript SDK uses Amazon Bedrock by default, so you must configure AWS credentials or pass another model.
- Strands Agents' TypeScript SDK requires a Node.js 22+ ESM project.
- Strands Agents marks some features as experimental, such as the OpenAI Live provider.
- Strands Agents documents bidirectional streaming for Python only.

## Usage data (the week of Oct 25–31, 2026)

- Board rank: #204 overall, #21 in Agent frameworks, momentum score 44
- Downloads that week: 8.9M (+1% vs 4-week average)
- New GitHub stars per week (last 12 weeks): Aug 9: 68; Aug 16: 70; Aug 23: 70; Aug 30: 114; Sep 6: 67; Sep 13: 141; Sep 20: 1.1k; Sep 27: 183; Oct 4: 127; Oct 11: 96; Oct 18: 98; Oct 25: 76

## Strands Agents alternatives

| Rank | Tool | Stars | Downloads/wk |
|---|---|---|---|
| #1 | [Harness²](https://index.devtoolwatch.com/tools/google-research-harness2/) | 21.7k | – |
| #2 | [HyperFrames](https://index.devtoolwatch.com/tools/heygen-com-hyperframes/) | 76.6k | 1.5M |
| #3 | [AgentScope](https://index.devtoolwatch.com/tools/agentscope-ai-agentscope/) | 35k | 51.4k |
| #4 | [AG-UI](https://index.devtoolwatch.com/tools/ag-ui/) | 16.9k | 4.6M |
| #5 | [LangGraph](https://index.devtoolwatch.com/tools/langgraph/) | 44.2k | 18.5M |
| #6 | [Bifrost](https://index.devtoolwatch.com/tools/maximhq-bifrost/) | 9.3k | – |

## Strands Agents FAQ

### Is Strands Agents free and open source?

Yes. Strands Agents's source code is on GitHub (strands-agents/harness-sdk) under the Apache 2.0 license, so you can use it for free.

### How do I install Strands Agents?

Run `pip install strands-harness`. The strands-agents package needs Python 3.10 or newer. The install section above has the other options from the project's README.

### Is Strands Agents still maintained?

Yes. The strands-agents/harness-sdk repository was last updated on Oct 25, 2026, and the latest release, harness-python/v0.3.2 on GitHub, came out on Oct 22, 2026.

### How popular is Strands Agents?

Strands Agents has 9k GitHub stars as of Nov 1, 2026. It gained 76 new stars in the week of Oct 25–31, 2026, against a 12-week average of 184 a week. Its main package had 8.9M downloads that week, +1% against its own 4-week average. That made it #21 among AI agent frameworks on the Devtoolwatch Board.

### Strands Agents vs LangGraph: how do they compare?

Strands Agents and LangGraph are both tracked by Devtoolwatch as AI agent frameworks. Strands Agents has 9k GitHub stars and LangGraph has 44.2k. In the week of Oct 25–31, 2026, Strands Agents's main package had 8.9M downloads and LangGraph's had 18.5M. LangGraph is a low-level orchestration framework and runtime for building, managing, and deploying long-running, stateful agents in Python or JavaScript. [LangGraph review](https://index.devtoolwatch.com/tools/langgraph/)

### Strands Agents vs Pydantic AI: how do they compare?

Strands Agents and Pydantic AI are both tracked by Devtoolwatch as AI agent frameworks. Strands Agents has 9k GitHub stars and Pydantic AI has 21k. In the week of Oct 25–31, 2026, Strands Agents's main package had 8.9M downloads and Pydantic AI's had 1.7M. Pydantic AI is a typed Python agent framework where models are swapped with a string, with tools, structured outputs, durable execution and realtime voice. [Pydantic AI review](https://index.devtoolwatch.com/tools/pydantic-ai/)

### What are the best alternatives to Strands Agents?

Other AI agent frameworks on the Devtoolwatch Board in the week of Oct 25–31, 2026, by momentum: Harness² (#1), HyperFrames (#2), AgentScope (#3), AG-UI (#4), LangGraph (#5). The alternatives table above has their numbers.

### What is Strands Agents?

Strands Agents is an open-source SDK for building and running AI agents in Python and TypeScript. It runs in your process with no hosted control plane. It covers lifecycle controls, tools, MCP, multi-agent patterns, memory, sessions, streaming, guardrails, tracing, and evals.

### What is the difference between Strands harness and the Strands Harness SDK?

Strands harness is a fully assembled agent created with one create_harness() or createHarness() call, with defaults for model, tools, memory, sessions, and context. The SDK lets you control the loop, tools, providers, memory, sessions, and hooks yourself.

### Which models does Strands Agents support?

Strands Agents has first-class support for Amazon Bedrock, Anthropic, OpenAI, and Gemini. It also supports providers such as Ollama, LiteLLM, llama.cpp, MistralAI, SageMaker, Vercel, and Writer, plus custom providers.

### Does Strands Agents support MCP?

Yes. Strands Agents includes MCP tool support. The project also ships an MCP server that gives coding agents knowledge about building with Strands Agents.

### Can Strands Agents be used with AI coding agents?

Yes. Strands Agents offers a plugin for Claude Code, Codex, and Cursor. The plugin adds scaffolding, tool authoring, MCP server connections, web framework guides, and LangGraph migration guidance.

### Does Strands Agents need a hosted service?

No. Strands Agents runs in your process with no hosted control plane. You still need access to a model provider, such as Amazon Bedrock or another supported provider.

## Sources

1. [strands-agents/harness-sdk on GitHub](https://github.com/strands-agents/harness-sdk)
2. [Strands Agents README](https://github.com/strands-agents/harness-sdk/blob/main/README.md)
3. [strands-agents on PyPI](https://pypi.org/project/strands-agents/)
4. [Strands Agents website](http://strandsagents.com/)
5. [Release harness-python/v0.3.2 (Oct 22, 2026)](https://github.com/strands-agents/harness-sdk/releases/tag/harness-python/v0.3.2)
6. [Release harness-cli/v0.2.2 (Oct 22, 2026)](https://github.com/strands-agents/harness-sdk/releases/tag/harness-cli/v0.2.2)
7. [Release harness-typescript/v0.3.1 (Oct 21, 2026)](https://github.com/strands-agents/harness-sdk/releases/tag/harness-typescript/v0.3.1)
8. [Hacker News: Strands Decider 2B: a small, open-source, decision model](https://news.ycombinator.com/item?id=49987076)
9. [Hacker News: Strands Harness](https://news.ycombinator.com/item?id=49817289)
10. [Hacker News: Strands Box: The Big Picture](https://news.ycombinator.com/item?id=49997144)
11. [Devtoolwatch Board, the week of Oct 25–31, 2026](https://index.devtoolwatch.com/)