This started with experiments with LangGraph. It then evolved into other LLM agentic experiments.
-
Exp1-Simple LangGraph example
https://github.com/vbookshelf/LangGraph-Experiments/tree/main/Exp1-Simple%20LangGraph%20example#exp1-simple-langgraph-example -
Exp2 - My LangGraph workflow
https://github.com/vbookshelf/LangGraph-Experiments/tree/main/Exp2-My%20LangGraph%20workflow -
Exp3 - Build an agent that runs the ReAct pattern using my LangGraph workflow
https://github.com/vbookshelf/LangGraph-Experiments/tree/main/Exp3-ReAct%20pattern%20w%20my%20langgraph%20workflow -
Exp4 - Use keyword generation instead of a ReAct pattern (re-implement exp3)
(Good example of my custom LangGraph workflow)
https://github.com/vbookshelf/LangGraph-Experiments/tree/main/Exp4-Use%20keyword%20generation%20instead%20of%20a%20ReAct%20pattern -
Exp5 - Create the exp2 email responder system without using LangGraph
(My agentt workflow)
https://github.com/vbookshelf/LangGraph-Experiments/tree/main/Exp5-Create%20exp2%20email%20responder%20without%20using%20LangGraph -
Exp6 - Create ReAct pattern using my agentic function workflow
(Dog weights and calculations example)
https://github.com/vbookshelf/LangGraph-Experiments/tree/main/Exp6-Create%20ReAct%20pattern%20using%20my%20agentic%20function%20workflow -
Exp7 - Adding chat to exp6 React pattern using my agentic functions
(Has a basic chat interface in the notebook - python user input with "while True" loop)
https://github.com/vbookshelf/LangGraph-Experiments/tree/main/Exp7-Adding%20chat%20to%20exp6%20React%20pattern%20using%20my%20agentic%20functions -
Exp8 - Create ReAct agent research assistant using gemini 1.5 flash api
https://github.com/vbookshelf/LangGraph-Experiments/tree/main/Exp8-Create%20ReAct%20agent%20using%20gemini%20flash%20api -
Exp9 - Semi autonomous ReAct research assistant with chat interface
(Uses the Gemini 1.5 Flash API)
https://github.com/vbookshelf/LangGraph-Experiments/tree/main/Exp9-Semi%20autonomous%20ReAct%20research%20assistant%20with%20chat%20interface -
Exp10 - Crate a code generation agent system
https://github.com/vbookshelf/LangGraph-Experiments/tree/main/Exp10-Code%20generation%20agent%20system -
Exp11 - Python - Create chat memory using summarization
https://github.com/vbookshelf/LangGraph-Experiments/tree/main/Exp11-%20Python%20-%20Create%20chat%20memory%20using%20summarization -
Exp12 - JS - Basic javascript chat app with summarization memory
https://github.com/vbookshelf/LangGraph-Experiments/tree/main/Exp12-%20JS%20-%20Basic%20javascript%20chat%20app%20with%20summarization%20memory -
Exp13 - Create a simple Mixture of Agents (MoA) implementation using Groq
https://github.com/vbookshelf/LangGraph-Experiments/tree/main/Exp13-Simple-MoA-implementation-using-Groq -
Exp14 - Create multi agent group chat with state dict memory
(The user chats with two agents at the same time. Written in Python.)
https://github.com/vbookshelf/LangGraph-Experiments/tree/main/Exp14%20-%20Rev2_multi_agent_group_chat_with_state_dict_memory -
Exp15 - Create a JS web app where the user chats with two agents at the same time
(Panel discussion: Are we living in a simulation?)
https://github.com/vbookshelf/LangGraph-Experiments/tree/main/Exp15%20-%20JS%20app%20with%20two%20agent%20chat -
Exp16 - Simple Pyodide example to run python code in the browser
https://github.com/vbookshelf/LangGraph-Experiments/tree/main/Exp16%20-%20Simple%20Pyodide%20example%20to%20run%20python%20code%20in%20the%20browser -
Exp17 - T-Bot PHP Thai and Spanish Practice Audio Chatbot
(Similar to the E-Bot English Practice Chatbot)
https://github.com/vbookshelf/LangGraph-Experiments/tree/main/Exp17%20-%20T-Bot%20PHP%20Thai%20and%20Spanish%20Practice%20Audio%20Chatbot
- It's easy to build simple agentic workflows without using LangChain or LangGraph (see exp5).
- It's a lot less code when LangGraph is not used.
- In a ReAct workflow the model can sometimes try to call all functions, that it needs to answer a question, at the sme time. It outputs a list of functions to be called.