Features
Multi-agent orchestration for deep research
Long-running task execution with planning and sub-tasking
Long-term and short-term memory for context
Docker-based sandbox with Browser, Shell, File, MCP, VSCode Server
Progressive loading of Agent Skills
Custom skill files via SKILL.md
Multi-model support: Doubao, DeepSeek, OpenAI, Gemini
File persistence and management within sandbox
Code generation and execution
Video generation from text/scenes
Image generation from text/scenes
Web scraping and data collection
Summarization and report generation
Subagents for complex tasks
Self-hosted with full data control
@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
