#3737th of 70 in Agent frameworks and SDKs
harness-sdk
Python and TypeScript SDK for building agents, with a prebuilt harness
Documentation ↗
(opens in a new tab)Website ↗
(opens in a new tab)Repository on GitHub ↗
(opens in a new tab)
- Stars
- 8.9k
- License
- Apache-2.0
- Last commit
- Oct 2026
- Last release
- Oct 2026
- Language
- TypeScript
5,419 on 1 Apr 20268,900 on 21 Oct 2026
From GitHub's daily star counts.
Checked on GitHub.
Overview
Monorepo for the Strands Agents SDKs in Python (3.10+) and TypeScript (Node.js 22+), plus a prebuilt harness created with create_harness() or createHarness(). The SDK covers the agent loop with turn and token limits, tools, MCP, structured output, multi-agent patterns, sessions, guardrails, tracing and evals. It runs in your process with no hosted control plane.
Who it is for: Developers who would otherwise hand-write an agent loop in Python or TypeScript
Strengths
- Same agent model in Python and TypeScript, with a CLI and an MCP docs server
- Providers include Bedrock, Anthropic, OpenAI, Gemini and Ollama, plus custom ones
- Hooks can intercept any step of the loop; tracing is on by default
- Harness gives a working agent in one call, and you can drop to the SDK
Weaknesses
- TypeScript Agent defaults to Amazon Bedrock, so AWS credentials are needed unless you set a model
- Several packages (harness, SDK, tools, CLI) to choose between; the split takes time to learn
- Harness defaults are described as benchmarked, but the README gives no numbers
- Resource requirements are not stated in the README
What it needs
- no GPU
- Needs Python 3.10+ or Node.js 22+, Model provider credentials (Amazon Bedrock is the TypeScript SDK default)
- Models: Amazon Bedrock, Anthropic, OpenAI, Gemini, Ollama
Also in Agent frameworks and SDKs
See all 70| Rank | Project | Score |
|---|---|---|
| 1 | ppt-masterAgent skill that turns documents into natively editable PowerPoint decks | 94 out of 100 |
| 2 | langchainPython framework for building LLM agents and apps from interoperable components | 89 out of 100 |
| 3 | browser-usePython library that lets an LLM agent control a web browser | 89 out of 100 |
| #4 | openai-agents-pythonPython SDK for multi-agent workflows with handoffs, guardrails and tracing | 89 out of 100 |
| #5 | mastraTypeScript framework for building AI agents and workflows | 89 out of 100 |