whereismyai
Agent Squad
Agent Squad is an open-source framework for routing user queries across multiple specialized AI agents while preserving context.
Agent Squad is an open-source framework designed to orchestrate teams of specialized AI agents. It works by analyzing user input with a classifier, matching the request to the appropriate agent using agent descriptions and conversation history, executing tool calls, and persisting the dialogue context. The framework supports pre-built integrations for Amazon Bedrock, Anthropic, OpenAI, AWS Lambda, and Amazon Lex, along with native functions, HTTP tools, and MCP servers. Available in Python, TypeScript, and Swift, it enables deployment on cloud infrastructure as well as on-device execution on iOS and macOS.
Agent Squad at a glance
Key features
- Context-aware query routing using classifiers or decision models
- Grounded two-LLM pattern separating data gathering from response presentation
- SupervisorAgent architecture for coordinating parallel sub-agent teams
- Streaming and non-streaming agent response handling
- On-device execution in Swift with realtime voice and MCP tool support
- Local-first chat storage with SwiftData or JSON files
- First-class tracing via OSLog and OTLP export to Langfuse, LangSmith, and Datadog
- Pre-built agents for Bedrock, Anthropic, OpenAI, Lex, and Lambda
Use it for
- Routing customer support requests to specialized agents with human-in-the-loop verification
- Building on-device conversational assistants with realtime voice for iOS and macOS
- Deploying multi-agent web applications using FastAPI, Streamlit, or Chainlit
- Generating grounded responses for sensitive data like pricing, stock status, and account balances
Alternatives to Agent Squad
- Composio — Equips AI agents with tools, managed user auth, and sandboxes across 1,500+ apps
- OpenViking — Unifies agent memory, knowledge RAG, and skills into a single directory structure.
- Beeuni — Agents retain context and improve skills over time through a controlled review process.
- Agentloop — Provides a clean agent loop with provider-agnostic LLM access, hooks, and safety budgets.
- Raven — Provides a persistent, self-evolving environment for all-domain agent collaboration.
- Alook — Connects existing local coding agents to team communication channels.
- OpenSquilla — Maximizes AI intelligence density while minimizing token costs through smart routing.
- Agenta — Build and automate AI agents that run on your own infrastructure or the cloud.
- Hybro — Orchestrates local and remote agents using the open Agent2Agent protocol.
- Rakazo — Complete ownership of data, models, and execution environment.
- AgentDock — It coordinates multi-device workflows and provides secure environment access.
- APIStack — Connect AI agents to over 100 tools using one API key or a single MCP endpoint.
All Agent Squad alternatives