#4747th of 70 in Agent frameworks and SDKs
agent-squad
Routes queries to specialized agents in Python, TypeScript and Swift
- Stars
- 7.8k
- License
- Apache-2.0
- Last commit
- Oct 2026
- Last release
- Sep 2026
- Language
- Python
7,476 on 1 Apr 20267,794 on 21 Oct 2026
From GitHub's daily star counts.
Checked on GitHub.
Overview
Agent Squad classifies each user query, picks the most suitable agent from those you register, and keeps conversation context across agents and sessions. It ships pre-built agents (Bedrock, Anthropic, OpenAI, Lex, Lambda), classifiers, storage, a SupervisorAgent for parallel team coordination, and a GroundedAgent that separates tool calling from answer writing. Python and TypeScript have feature parity; a Swift runtime targets iOS 16+ and macOS 14+.
Who it is for: Developers building multi-agent apps in Python, TypeScript or Swift
Strengths
- Same orchestration model in Python 3.11+, TypeScript and on-device Swift
- Classifier routing, streaming and non-streaming responses, pluggable storage and retrievers
- SupervisorAgent queries sub-agents in parallel; GroundedAgent limits replies to fetched tool output
- Swift runtime supports MCP tools, realtime voice and OTLP tracing
Weaknesses
- Pre-built agents and examples lean toward AWS (Bedrock, Lex, Lambda)
- JevClassifier needs a TypeSafe AI API key and a paid external service
- Swift quick start pins the main branch, not a tagged version
- Moved from awslabs and renamed from multi-agent-orchestrator; old links and dependencies break
What it needs
- no GPU
- Needs LLM provider API (Bedrock, Anthropic or OpenAI), TypeSafe AI API key (JevClassifier only)
- Models: Amazon Bedrock, Anthropic, OpenAI, Amazon Lex, AWS Lambda
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 |