Lightweight & Asynchronous
Engineered from the ground up on standard Python asyncio to handle high-concurrency conversational turns with low overhead.
chatflow-agentOpen Source
About chatflow-agent
A lightweight, asynchronous Python multi-agent framework designed for building conversational workflows with native handoffs and multi-channel connectivity.
Origin & Purpose
chatflow-agent was created to address a clear gap in conversational AI development: connecting LLMs to real-world communication channels without introducing bloated, monolithic orchestration layers.
Instead of imposing complex distributed systems, chatflow-agent focuses on pure Python simplicity, native asynchronous concurrency via asyncio, and reliable SQLite-backed session persistence with turn rollback safety.
Core Architectural Principles
Native Multi-Agent Handoffs
Orchestrate context-aware transitions between specialized agents seamlessly, preserving history and conversational context.
Real-World Channels
Production-ready connectors for WhatsApp, Telegram, Webhooks with HMAC signature verification, and command-line interfaces.
State & Memory Integrity
Built-in SQLite session persistence with strict sequential turn processing to eliminate race conditions and corrupted histories.
Project Leadership
Creator & Lead Maintainer
Gustavo Baranda
Architect and lead maintainer of chatflow-agent. Designed the core async multi-agent runtime, native handoff mechanisms, and channel adapters to make robust conversational AI accessible to developers worldwide.
Open Source
Global Community & Open Source
chatflow-agent thrives thanks to contributions from the global open-source community. Developers, translators, testers, and users worldwide help improve the framework with every release.
Ready to Build Conversational Multi-Agent Systems?
Explore our comprehensive documentation or join discussions with the developer community on GitHub.