PraisonAI

PraisonAI describes itself as a production-ready multi-AI-agents framework with self-reflection, offering a low-code way to create agents that automate tasks from simple to complex. It ships as a Python SDK (`pip install praisonaiagents` for the core, `pip install praisonai` for the full package), a JavaScript SDK (`npm install praisonai`) and a no-code CLI, with a one-minute quick start to a first agent.

The feature tables cover single, multi and auto-generated agents, self-reflection, reasoning and multimodal agents, and workflow patterns such as routing, parallel execution, looping over a list or CSV, an evaluator-optimiser repeat loop, conditional steps, branching and early stop. The Python examples also include planning mode, deep research and query-rewriter agents, zero-dependency agent memory, auto-generated memories, hooks, shadow git checkpoints, background tasks, a policy engine, thinking budgets, context compaction, and the MCP and A2A (agent-to-agent) protocols. Many model providers are supported. A performance table in the README claims the fastest agent instantiation time among the frameworks it tests; it is the project's own benchmark, and the repository is MIT licensed.