What it is
The LangChain team’s agent orchestration framework models flow as a stateful graph: nodes are steps, edges are transition conditions, with loops, branches and human-in-the-loop support. It tackles the problem of agent tasks being hard to control.
Highlights
- Flows modelled as stateful graphs with explicit loops and branches
- Supports interruption, resume and human intervention
- Execution state can persist for long-running tasks
- Integrates seamlessly with the LangChain ecosystem
How to use
Install via pip, define nodes and edges to build a graph, then add checkpointing to persist state for long jobs.
License
Released under MIT. Read the terms before commercial use or redistribution, especially if you plan to offer it as a service.