Project 30: AI Support Agent Capstone

AI Engineer

Current Task

Objective

Capstone: wire a prompt, a retrieval chain, a memory-backed conversation, and a tool-using agent into one pipeline — the complete LangChain workflow in one script.

Task: assemble all of it, run one turn, and print the memory length and the final answer.

index.py

from langchain.memory import ConversationBufferMemory
from langchain.chains import create_retrieval_chain
from langchain_core.tools import tool

memory = ConversationBufferMemory()

@tool
def flag_for_review(item_id: str) -> str:
    """Flag an admin dashboard item for manual review."""
    return "flagged"

retrieval_chain = create_retrieval_chain(retriever, question_answer_chain)
response = retrieval_chain.invoke({"input": "Why was item #482 flagged?"})

memory.save_context({"input": "Why was item #482 flagged?"}, {"output": response["answer"]})

print(len(memory.chat_memory.messages))
print(response["answer"])

* Hint: Correct characters turn green, incorrect ones turn red.

Live Preview

🖼️

Verify your code to see the preview

← Previous