LangChain

What it is

LangChain is the framework that turns a raw LLM API call into a real agent — chaining prompts, calling tools, retrieving from vector stores, holding memory across turns. It's the connective tissue of every modern AI worker.

How Vaaani uses it

• Building agents that call your APIs (Stripe, HubSpot, internal tools)
• RAG pipelines over your documents with retrievers and rerankers
• Conversational memory across long sessions with summarization
• Streaming responses to React frontends via LangChain.js

Why it makes the cut

Every chatbot Vaaani ships uses LangChain (or LangGraph) at its core. It's the difference between a 100-line prompt hack and a maintainable agent.

Sample code

from langchain.agents import create_react_agent
from langchain_openai import ChatOpenAI

llm = ChatOpenAI(model="gpt-4o")
agent = create_react_agent(llm, tools=[search, db_lookup])

agent.invoke({"input": "Find Q2 revenue from the dashboard"})