fast-agent and jaf-py
About fast-agent
evalstate/fast-agent
Code, Build and Evaluate agents - excellent Model and Skills/MCP/ACP Support
Supports multiple LLM providers (Anthropic, OpenAI, Google, Azure, Ollama, Deepseek) with native MCP server integration via stdio and HTTP transports, plus structured outputs, vision, and PDF capabilities. CLI-first architecture with prompt_toolkit TUI, live streaming responses, and diagnostic tools for inspecting MCP transport behavior. Enables declarative agent composition through simple Python decorators and YAML configs, with built-in skills registry, shell mode, and ACP client compatibility for embedding agents into other applications.
About jaf-py
xynehq/jaf-py
Functional Python agent framework with MCP support, enterprise security, immutable state, and production-ready observability for building scalable AI systems.
Built on purely functional architecture with immutable state management and composable tools, JAF enforces type safety through Pydantic and achieves feature parity with its TypeScript counterpart. The framework integrates MCP servers via stdio and SSE transports, enables hierarchical agent-as-tool composition, and provides enterprise-grade observability through OpenTelemetry and Langfuse tracing alongside built-in input/output guardrails and RBAC. FastAPI server, Redis/PostgreSQL memory providers, and visual architecture diagrams (Graphviz) round out production deployment capabilities.
Related comparisons
Scores updated daily from GitHub, PyPI, and npm data. How scores work