# 2FastLabs/agent-squad

Flexible and powerful framework for managing multiple AI agents and handling complex conversations

Repository: https://github.com/2FastLabs/agent-squad
Canonical: https://ross.abutalabs.com/products/agent-squad
Homepage: https://2fastlabs.github.io/agent-squad/
Language: Swift
License: Apache-2.0
License Family: permissive
Topics: aws, aws-bedrock, anthropic-claude, aws-cdk, aws-lambda, chatbot, generative-ai, machine-learning, openai, serverless, python, typescript, agents, ai-agents, ai-agents-framework, openaiapi, framework, orchestrator, anthropic, agentic-ai
Last push: 2026-08-15T23:27:30+00:00

## Health v2 (maintenance only)
Score: 87/100 (v2, computed 2026-09-03T02:20:16.233290+00:00)
- activity 97, release rhythm 93, longevity 55
- inputs: {"age_days": 771, "days_push": 18, "days_rel": 50, "gap_med": 5, "n_releases_24m": 26}
- flags: none
- formula: round(0.45*activity + 0.35*rhythm + 0.20*longevity); archived -> min(score, 10)

## Adoption (not part of the score)
Stars 7748, forks 740 (observed 2026-08-28T04:10:04.136838+00:00)

## What it is
Agent Squad is an open-source framework for orchestrating multiple AI agents, routing each user query to the most suitable specialized agent via intent classification while maintaining conversation context. It offers Python, TypeScript, and Swift runtimes, with pre-built agents for Bedrock, Anthropic, OpenAI, Lex, and Lambda, plus extensible interfaces for custom agents, classifiers, storage, and retrievers.

## Use cases
- route user queries to specialized AI agents automatically
- build a multi-agent chatbot with shared conversation context
- orchestrate agents across AWS Bedrock, Anthropic, and OpenAI
- run agent orchestration on-device on iOS or macOS with Swift
- add MCP tools and realtime voice to an agent pipeline
- deploy a serverless multi-agent assistant on AWS Lambda
- stream agent responses in a chat application

## When to choose
- you need multiple specialized LLM agents behind a single conversational interface
- you want pluggable classifiers, storage backends, and retrievers without vendor lock-in
- you're deploying on AWS (Lambda, Bedrock, DynamoDB) or need Python/TypeScript parity
- you want on-device agent orchestration with MCP tools and voice on Apple platforms

## When to avoid
- you need a single simple LLM call with no routing or agent logic
- you require a full agent workflow/autonomy engine with long-running task planning
- your stack doesn't include Python, TypeScript/Node.js, or Swift

## Facets
- artifact type: framework
- maturity: active
- function: agent-framework, chatbot, llm-inference, rag, mcp, sdk
- domain: large-language-models, chatbots, developer-tools, cross-platform
- platform: python, serverless, cross-platform
- tags: multi-agent-orchestration, intent-classification, aws-bedrock, anthropic, openai, streaming, conversation-management, on-device-ai, ai-agents, nodejs, swift, ios, macos

## Member repositories
- 2FastLabs/agent-squad (main) score 87

## Provenance
- Observed fields: from GitHub, fetched 2026-08-28T04:10:04.136838+00:00.
- Health v2: computed from the inputs above; adoption is never an input.
- Inferred fields (summary, facets, guidance): AI-extracted, prompt v1, taxonomy v1, on 2026-08-29T17:35:09.544449+00:00, confidence not recorded.
  - readme: https://github.com/2FastLabs/agent-squad (fetched 2026-08-28T04:10:04.136838+00:00, sha 8fe79ddc93ba)
  - homepage: https://2fastlabs.github.io/agent-squad/ (fetched 2026-08-29T08:31:09.769028+00:00, sha 1b38160221c2)
- Data as of 2026-08-30T08:39:29.467469+00:00.