📚 DevOps Documentation Chatbot (RAG + Gemini + FAISS)

A lightweight AI-powered documentation chatbot built using Flask, LangChain, Google Gemini, and FAISS.

The system uses Retrieval-Augmented Generation (RAG) to answer questions based on your own documents. Instead of hallucinating answers, it retrieves relevant document chunks and sends them to Gemini for grounded responses.

Currently the chatbot indexes DevOps documentation such as:

• Docker
• CI/CD

But it can be extended to any domain knowledge base.

---

---

🚀 Features

• Retrieval-Augmented Generation (RAG) architecture • Google Gemini LLM for responses • Gemini embedding model for semantic search • FAISS vector database for fast retrieval • Session-based conversation memory • Flask API backend • Simple web UI (index.html) • Source document references returned with answers • Rate-limit safe embedding pipeline for free-tier Gemini API

---

🧠 Architecture

User Question
      │
      ▼
Flask API (/chat)
      │
      ▼
History-aware Retriever
      │
      ▼
FAISS Vector Store
      │
Retrieve Top-K Relevant Chunks
      │
      ▼
Gemini LLM (RAG Prompt)
      │
      ▼
Answer + Sources

This prevents hallucinations by forcing the model to use retrieved context from your docs.

---

📂 Project Structure

project/
│
├── docs/                     # Knowledge base documents
│   ├── docker.txt
│   └── cicid.txt
│
├── vector_store/             # Generated FAISS index
│
├── embed_docs.py             # Creates embeddings from d
├── main.py                   # Flask RAG chatbot API
├── index.html                # Frontend chat UI
├── requirements.txt          # Python dependencies
│
└── README.md

---

⚙️ Installation

1️⃣ Clone the repository

git clone <repo-url>
cd <project-folder>

---

2️⃣ Create virtual environment

python3 -m venv venv
source venv/bin/activate

Windows:

venv\Scripts\activate

---

3️⃣ Install dependencies

pip install -r requirements.txt

---

🔑 Setup Gemini API Key

Get an API key from:

https://ai.google.dev/

Then export it as an environment variable.

ENV:

GEMINI_API_KEY="your_api_key_here"

---

📄 Add Documentation

Place .txt files inside the docs/ folder.

Example:

docs/
   docker.txt
   cicid.txt

You can add any domain documents such as:

• Kubernetes
• Terraform
• Jenkins
• DevOps Interview Notes
• Internal Company Docs

---

🔢 Generate Vector Embeddings

Run the embedding script:

python embed_docs.py

What this script does:

1. Loads documents from docs/
2. Splits them into chunks
3. Generates embeddings using Gemini Embedding Model
4. Stores vectors in FAISS
5. Saves index in vector_store/

Example output:

Loading docs …
Loaded 2 file(s)
Split into 34 chunks
Embedding with gemini-embedding-001 …
Vector store saved to 'vector_store/'
Total vectors indexed: 34

---

▶️ Run the Chatbot

Start the Flask server:

python main.py

Server runs at:

http://localhost:5000