RAG TERMINAL

A fully local AI chatbot with document upload and RAG. No API keys. No internet. Runs entirely on your machine.

What it does

Upload a PDF or TXT file and ask questions about it. The chatbot retrieves relevant sections from your document and uses Gemma 2B (via Ollama) to answer — all locally, all private.

Without a document, it works as a general-purpose AI assistant.

---

Stack

Layer	Tool
LLM	Gemma 2B · Ollama
Backend	Python · FastAPI · LangChain
Vector DB	ChromaDB
Embeddings	sentence-transformers
Frontend	HTML · CSS · JS
Deploy	Docker · docker-compose

---

Setup

Full walkthrough → SETUP.txt

# 1. Pull the model
ollama pull gemma:2b

# 2. Create environment
cd backend
python -m venv venv
.\venv\Scripts\Activate.ps1   # Windows
source venv/bin/activate       # Mac/Linux

# 3. Install and run
pip install -r requirements.txt
uvicorn main:app --host 0.0.0.0 --port 8000 --reload

Open http://localhost:8000

Docker alternative:

docker-compose up --build

---

Project structure

rag-chatbot/
├── backend/
│   ├── main.py            API server
│   ├── rag.py             RAG pipeline
│   ├── requirements.txt
│   └── Dockerfile
├── frontend/
│   └── index.html         Chat UI
├── chromadb_data/         Auto-created · stores embeddings
├── uploads/               Auto-created · stores your files
├── docker-compose.yml
├── README.md
└── SETUP.txt              Full setup guide

---

Common errors

Error	Fix
Connection refused :11434	ollama serve
model not found	ollama pull gemma:2b
Activate.ps1 blocked	Set-ExecutionPolicy RemoteSigned -Scope CurrentUser
Slow responses	Normal on CPU — just wait

---

Built for AI Advanced Skill Development Lab