Features
Open-source Python framework for building agents
Model-driven orchestration; no rigid workflows
Built-in web search and web browsing tools
Code execution: local, Docker, E2B sandbox
MCP client for model-context protocol integration
Document input and output handling
Skills system for modular domain-specific instructions
Flexible tool execution via generic Tool class
Context management with automatic conversation summarization
Multi-provider LLM support (OpenAI-compatible, LiteLLM, custom)
Multimodal support: images, video, audio
Session-based execution with lifecycle management
Built-in logging and file output handling
OpenResponses client for OpenAI Responses API
Sub-agents capability and caching support
@ai_fn decorator for LLM-powered Python functions
@ai_classifier decorator for text classification
Structured data extraction via Pydantic models
Agent loops with tool calling and function calling
Built-in streaming (SSE) support
Async-first API for concurrent applications
Rate limiting and automatic retries
LLM response caching for performance
Concurrency control for workload management
CLI tool for monitoring and debugging
SQLite state store for persistence
OpenAI and Anthropic model support
Embeddings generation for semantic search
Local execution with no cloud dependency
Self-hosted deployable as a library
