Features
Multi-agent workflows: Sequential, Parallel, DAG, Loop + subworkflows
LLM-agnostic: supports multiple model providers
MCP protocol for dynamic tool discovery and integration
Built-in vector databases: pgvector, Weaviate for RAG
Real-time streaming with 13 chunk types
Typed errors and comprehensive retry logic
Memory management and knowledge base capabilities
State sharing across agents and subworkflows
Compiled binaries for single-binary deployment
Full test coverage and production readiness
Migration guide from deprecated Legacy APIs
8 core builder methods (streamlined API)
Goroutine-based true parallel agent execution
Open source under Apache 2.0 license
Multi-agent orchestration with handoffs
Sandbox Agents for containerized long-running tasks
Agent-as-tool delegation (v0.15.0+)
Realtime Agents with gpt-realtime-2 voice support (v0.17.6+)
Human-in-the-loop mechanisms
Automatic session history management
Built-in tracing for debugging
Provider-agnostic LLM support (100+ models via LiteLLM)
Redis session support (optional, v0.17.6+)
Instructions, tools, and guardrails configuration
Jupyter notebook compatibility
Supports OpenAI Responses and Chat Completions APIs
