Discord-RAG-FAQ-Chatbot

A backend system that connects a Discord chatbot with a Retrieval-Augmented Generation (RAG) agent powered by Google Gemini.
The backend is responsible for API design, LLM integration, logging, observability, and containerized deployment.

This project focuses on backend architecture and operational excellence, rather than frontend UI or model training.

Architecture Overview

Discord Bot
-> HTTP (POST /api/rag-query)
-> FastAPI Backend (Docker)
-> Prompt-based RAG logic
-> Google Gemini API

Key Responsibilities of the Backend

• Receive user questions from a Discord bot
• Process Python-related queries using a Gemini-powered RAG agent
• Return structured responses
• Provide logging and observability
• Support containerized deployment via Docker

Running Locally (Without Docker)

Note: you need to create a Discord bot first in Discord Developer Portal.

1. Install dependencies

pip install -r requirements.txt

2. Fill Keys in .env

GEMINI_API_KEY=your_api_key
DISCORD_TOKEN=your_discord_bot_token

3. Start Backend Under the Project Folder in Terminal

uvicorn app.main:app --reload --env-file .env

4. Start Discord Bot Under the Project Folder in New Terminal

python app/discord_bot.py

5. Use in Discord

Type the question after !py to ask your bot in Discord.
Example:

!py What is a Python decorator?

Running with Docker

Note: you may need to download Docker Desktop first.

1. Fill Keys in .env

GEMINI_API_KEY=your_api_key
DISCORD_TOKEN=your_discord_bot_token

2. Build Image

After running Docker Desktop, open a terminal and navigate to the project folder to type the below command:

docker build -t discord-rag-backend .

3. Run Container

In the same terminal, run:

docker run --env-file .env -p 8000:8000 discord-rag-backend

4. Test API

Keep the last terminal open, and open new terminal under the project folder. Type:

curl -X POST http://localhost:8000/api/rag-query ^
  -H "Content-Type: application/json" ^
  -d "{\"user_id\":\"test\",\"query\":\"What is a Python decorator?\"}"

You should get answer.

5. Use in Discord

Open new terminal under the project folder and type:

python app/discord_bot.py

Open Discord and type the question after !py to ask your bot in Discord.