Skip to main content
Huzbi's room

Lang-Projects & Agent Experiments

A lab repo for LangChain and LangGraph experiments, agentic loops, and multi-agent workflows that mostly behave.

[MOD] AI & agents

The short version

A central hub for my LLM orchestration experiments, which is a polite way of saying a folder where LangChain and LangGraph ideas go to be tested. LangChain handles the linear chains and modular pieces; LangGraph handles the stateful, loopy, conditional parts where an agent gets to argue with itself. At the moment that means a standard RAG pipeline for loading documents as outside knowledge and a reflection agent that critiques and refines its own output.

What's in it

  • LangChain-Rag: a standard retrieval-augmented generation system that loads documents and uses them as an external source of knowledge.
  • LangGraph-Reflection-Agent: a reflection-style agent built with LangGraph that improves its output through self-critique and refinement.
  • A playground that starts at simple RAG and is allowed to wander into autonomous multi-agent territory, as the README puts it.

The stack

PythonLangChainLangGraphRAGAgents

Languages: Python. Agentic Pipeline Module.

AI Anime Image Generator

An interactive web app that turns a prompt into an anime character or scene, with enough knobs to keep tweaking the result.

Speech Translator

A real-time speech-to-speech and transcription pipeline that listens, transcribes, translates, then talks back in another language.

Keep looking

There are more artifacts on the bench.

The lab directory lists every project, and the GitHub log holds all 40 public repositories and the year's lit contribution windows.