From Local to Live: Deploying an Audio Transcription & Translation App Using Streamlit, Docker, and GitHub Actions

🧠 Introduction

In this blog post, I’ll walk you through how I built, containerized, and deployed a real-time audio transcription and translation app using:

• 🎧 Streamlit for the frontend

• 🤖 OpenAI Whisper API for transcription

• 🌐 Translation API for multilingual output

• ☁️ AWS S3 and EC2 for cloud deployment

• 🐳 Docker for containerization

• 🔄 GitHub Actions for CI/CD automation

Whether you're a Python developer, DevOps learner, or looking to automate your deployments — this article is packed with practical takeaways! 🚀

---

📌 Project Overview

GitHub Repository: S3-Audio-Transcription-Translation-App

This Streamlit app allows users to:

• Upload audio files (MP3/WAV)

• Transcribe audio using OpenAI’s Whisper

• Translate the transcript into multiple languages

• Store metadata in MySQL

• Upload audio files to AWS S3

---

🛠️ Technologies Used

---

🚧 Step-by-Step Development

1️⃣ Build the Streamlit App

I used Streamlit to build a clean UI that takes in audio files, triggers transcription, and displays translated text. Here's a snippet:

import streamlit as st

st.title("Audio Transcription & Translation App")
uploaded_file = st.file_uploader("Upload an audio file", type=["mp3", "wav"])

if uploaded_file:
    st.audio(uploaded_file)
    # Process transcription and translation...

---

2️⃣ Integrate OpenAI & Translation APIs

Once the file is uploaded, the backend sends it to OpenAI’s Whisper API and returns a transcript. This is then sent to a translation API (e.g., Google Translate or OpenAI GPT) based on user-selected language.

---

3️⃣ Containerize with Docker

To ensure consistency across development and production, I created a Dockerfile:

# Use a slim Python base image
FROM python:3.9-slim

RUN apt-get update && apt-get install -y \
    build-essential \
    libpq-dev \
    portaudio19-dev \
    libasound-dev \
    # Add other system dependencies as needed, e.g., libgfortran5 for numpy
    --no-install-recommends && \
    apt-get clean && rm -rf /var/lib/apt/lists/*

# Set the working directory
WORKDIR /app

# Copy the requirements file and install dependencies
COPY requirements.txt .
RUN pip install -r requirements.txt

# Copy the Streamlit application code
COPY . .

# Expose the default Streamlit port
EXPOSE 8501
# Command to run the Streamlit app with environment variables
# The actual environment variables will be passed during `docker run` on     EC2
CMD ["streamlit", "run", "appfouthjuly.py", "--server.port=8501", "--server.address=0.0.0.0"]

---

4️⃣ Automate with GitHub Actions

Every push to main triggers a GitHub Actions pipeline that:

• ✅ Runs unit tests using pytest

• 🐳 Builds Docker image

• 📤 Pushes it to Docker Hub

• 🚀 Deploys it to an EC2 instance via SSH

name: CI/CD Pipeline for Streamlit App to EC2 (Docker Hub)

on:
  push:
    branches:
      - main  # or your desired branch

jobs:
  build-and-deploy:
    runs-on: self-hosted

    steps:
      - name: Checkout code
        uses: actions/checkout@v4

      - name: Set up Docker Buildx # Add this step to set up Buildx
        uses: docker/setup-buildx-action@v3

      - name: Log in to Docker Hub
        uses: docker/login-action@v3
        with:
          username: ${{ secrets.DOCKERHUB_USERNAME }}
          password: ${{ secrets.DOCKERHUB_TOKEN }}

      - name: Run Unit Tests
        run: |
          pytest tests/ --maxfail=1 --disable-warnings --tb=short

      - name: Build and push Docker image to Docker Hub
        uses: docker/build-push-action@v5
        with:
          context: .
          push: true
          tags: ${{ secrets.DOCKERHUB_USERNAME }}/${{ secrets.DOCKERHUB_REPO }}:latest

      - name: Deploy to EC2
        uses: appleboy/ssh-action@master
        with:
          host: ${{ secrets.EC2_PUBLIC_DNS }}
          username: ${{ secrets.EC2_USERNAME }}
          key: ${{ secrets.EC2_SSH_PRIVATE_KEY }}
          script: |
            echo "Pulling latest image..."
            sudo docker pull ${{ secrets.DOCKERHUB_USERNAME }}/${{ secrets.DOCKERHUB_REPO }}:latest

            echo "Stopping/removing existing container if it exists..."
            if [ "$(sudo docker ps -aq -f name=my-streamlit-app)" ]; then
                sudo docker stop my-streamlit-app || true
                sudo docker rm my-streamlit-app || true
            fi

            echo "Checking if any other container is publishing port 8501..."
            if [ "$(sudo docker ps -q -f publish=8501)" ]; then
                CONTAINER_ID_USING_PORT=$(sudo docker ps -q -f publish=8501)
                echo "Stopping and removing container $CONTAINER_ID_USING_PORT using port 8501..."
                sudo docker stop $CONTAINER_ID_USING_PORT || true
                sudo docker rm $CONTAINER_ID_USING_PORT || true
            fi

            echo "Forcefully killing any process on port 8501 (just in case)..."
            sudo fuser -k 8501/tcp || true

            echo "Running new container..."
            sudo docker run -d \
              --network host \
              --name my-streamlit-app \
              -e MYSQL_HOST="127.0.0.1" \
              -e MYSQL_USER="${{ secrets.MYSQL_USERNAME }}" \
              -e MYSQL_PASSWORD="${{ secrets.MYSQL_PASSWORD }}" \
              -e MYSQL_DATABASE="${{ secrets.MYSQL_DATABASE_NAME }}" \
              -e OPENAI_API_KEY="${{ secrets.OPENAI_API_KEY }}" \
              -e S3_BUCKET_NAME="${{ secrets.S3_BUCKET_NAME }}" \
              -e AWS_ACCESS_KEY_ID="${{ secrets.AWS_ACCESS_KEY_ID }}" \
              -e AWS_SECRET_ACCESS_KEY="${{ secrets.AWS_SECRET_ACCESS_KEY }}" \
              -e AWS_REGION="${{ secrets.AWS_REGION }}" \
              ${{ secrets.DOCKERHUB_USERNAME }}/${{ secrets.DOCKERHUB_REPO }}:latest

---

5️⃣ Deploy to EC2

Using appleboy/ssh-action, the EC2 server pulls the latest image, removes existing containers (if any), and runs the app on port 8501.

 - name: Deploy to EC2
        uses: appleboy/ssh-action@master
        with:
          host: ${{ secrets.EC2_PUBLIC_DNS }}
          username: ${{ secrets.EC2_USERNAME }}
          key: ${{ secrets.EC2_SSH_PRIVATE_KEY }}
          script: |
            echo "Pulling latest image..."
            sudo docker pull ${{ secrets.DOCKERHUB_USERNAME }}/${{ secrets.DOCKERHUB_REPO }}:latest

            echo "Stopping/removing existing container if it exists..."
            if [ "$(sudo docker ps -aq -f name=my-streamlit-app)" ]; then
                sudo docker stop my-streamlit-app || true
                sudo docker rm my-streamlit-app || true
            fi

            echo "Checking if any other container is publishing port 8501..."
            if [ "$(sudo docker ps -q -f publish=8501)" ]; then
                CONTAINER_ID_USING_PORT=$(sudo docker ps -q -f publish=8501)
                echo "Stopping and removing container $CONTAINER_ID_USING_PORT using port 8501..."
                sudo docker stop $CONTAINER_ID_USING_PORT || true
                sudo docker rm $CONTAINER_ID_USING_PORT || true
            fi

            echo "Forcefully killing any process on port 8501 (just in case)..."
            sudo fuser -k 8501/tcp || true

---

🔐 Environment Variables

Secrets were handled using GitHub Secrets for:

• AWS keys

• OpenAI API key

• MySQL credentials

• S3 bucket info

---

🧪 Testing with Pytest

Before deploying, pytest is used to validate core functionality (e.g., upload, API call mocks):

- name: Run Unit Tests
        run: |
          pytest tests/ --maxfail=1 --disable-warnings --tb=short

---

✅ Final Outcome

The app is fully functional and accessible via browser after deployment. On every push to main, changes are automatically tested, built, and deployed — zero manual steps!

---

💡 What I Learned

• Writing production-ready Streamlit apps

• Docker fundamentals and image optimization

• CI/CD with GitHub Actions

• EC2 and SSH deployment strategies

• Managing secrets and environment variables securely

---

🔗 Project Links

• 🔸 GitHub Repo: kri-sh27/S3-Audio-Transcription-Translation-App

• 🔸 Live App: http://<your-ec2-public-ip>:8501

---

🙌 Conclusion

This was a rewarding full-stack + DevOps project that touched frontend, backend, cloud, and automation. I’ll continue improving the app by adding:

• 🎙 Real-time microphone support

• 📈 Logging and monitoring with Prometheus/Grafana

• 🔐 JWT Auth for user access control

Thanks for reading! Feel free to fork the repo, try the app, and connect with me if you have questions or ideas.

---

📬 Let's Connect!

• 🔗 GitHub

• 💼 LinkedIn