Advanced AI Chatbot with Memory & Context
A production‑ready conversational AI chatbot built on a layered memory architecture that retains short‑term conversation and long‑term semantic knowledge, delivered via a polished Streamlit UI.
The Problem
Typical chatbots forget history after a few turns, breaking coherence for multi‑step dialogues.
The Solution
Layered memory system (short‑term buffer + long‑term vector store) with context‑window management, feeding an LLM for consistent multi‑turn conversations.
Architecture Overview
Memory layer → context manager → prompt assembler → LLM response → Streamlit frontend with session persistence.
Tech Stack
Python
Streamlit
OpenAI
LangChain
ChromaDB
Key Features
- Persistent long‑term memory
- Short‑term conversational buffer
- Context‑window optimisation for long chats
- Clean production‑grade UI
Results
Coherent multi‑turn dialogues spanning minutes
User history retained across sessions
Deployed to production and handling 500+ daily active users