A Retrieval-Augmented Generation (RAG) chatbot built with TinyLlama via Ollama, designed to answer questions based on classic Marxist texts.

A practical demonstration of local RAG pipelines on domain-specific document collections — the same architecture used in enterprise knowledge management systems, applied here to political philosophy.

User query → ChromaDB vector search → relevant text chunks → TinyLlama (local, via Ollama) → answer with source context

• Loads .txt, .pdf, .docx versions of Marxist works
• Fully local — no API keys, no data leaves your machine
• Vector-based retrieval for contextually relevant answers
• Same RAG pattern as production enterprise systems

git clone https://github.com/joyboseroy/marxbot.git
cd marxbot
pip install -r requirements.txt

# Install Ollama from https://ollama.com
ollama run tinyllama

# Get texts from https://www.marxists.org/archive/marx/works/
# Place in marxtexts/ folder

python marxbot.py