LangGraphGo (lango) provides a Go implementation of LangGraph, enabling developers to build complex, stateful AI agent and workflow applications. It targets Go developers seeking robust, production-ready tools for LLM orchestration, offering parallel execution, persistence, advanced state management, and enhanced observability, aiming for feature parity with Python's LangGraph.

How It Works

This project implements a stateful, graph-based execution engine for LLM applications in Go. Its core approach features parallel node execution with thread-safe state merging, generic type-safe state management, and seamless integration with langchaingo. Key advantages include robust persistence options (Redis, Postgres, SQLite, File), advanced state manipulation via custom reducers and smart messages, and programmatic tool calling (PTC) for reduced latency and token usage.

Quick Start & Requirements

Installation is via go get github.com/smallnest/langgraphgo. Cloning requires handling Git submodules (--recurse-submodules or git submodule update --init --recursive). The project relies on a Go environment and integrates with langchaingo. The website is http://lango.rpcx.io.

Highlighted Details

• Parallel Execution: Supports concurrent node execution with thread-safe state merging.
• Persistence & Reliability: Offers checkpointers for Redis, Postgres, SQLite, and File, enabling state recovery and reliable execution.
• Programmatic Tool Calling (PTC): Enables LLMs to generate code that calls tools directly, achieving up to 10x reduction in latency and token usage.
• Developer Experience: Includes graph visualization (Mermaid, DOT, ASCII), built-in tracing and metrics for observability, and human-in-the-loop workflows for interactive control.
• Pre-built Agents: Provides factory functions for common agent patterns like ReAct, Supervisor, and Planning agents.

Maintenance & Community

The project welcomes contributions, suggesting feature issue creation before submitting pull requests. No specific community channels (like Discord or Slack) or roadmap links are provided in the README.

Licensing & Compatibility

The project is released under the MIT License. It is designed to work seamlessly with langchaingo and does not appear to have explicit restrictions for commercial use.

Limitations & Caveats

As a fork aiming for feature parity with Python's LangGraph, ongoing development and potential feature gaps compared to its Python counterpart may exist. The README does not detail specific known bugs, alpha status, or deprecations.