🎙️ AI Podcast Generator (CrewAI + Ollama + Gradio + Docker)

An end-to-end AI-powered podcast generator that:

• Scrapes blog content using Firecrawl
• Summarizes it using CrewAI agents + Ollama (DeepSeek-R1)
• Converts the summary into natural voice using ElevenLabs
• Provides a clean Gradio web interface
• Runs fully inside Docker with local Ollama LLM

---

🚀 Features

• ✅ Blog scraping with Firecrawl
• ✅ Multi-agent reasoning using CrewAI
• ✅ Local LLM inference using Ollama (no OpenAI/Gemini needed)
• ✅ High-quality text-to-speech using ElevenLabs
• ✅ Interactive Gradio web UI
• ✅ Fully Dockerized for easy deployment
• ✅ Secure .env based API key handling

---

🧠 Architecture Overview

User → Gradio UI → CrewAI Agents │ ├── Firecrawl (Scraping) ├── Ollama (DeepSeek-R1 LLM) └── ElevenLabs (Text-to-Speech)

---

🛠️ Tech Stack

• Python 3.11
• CrewAI
• Ollama (DeepSeek-R1)
• Firecrawl API
• ElevenLabs API
• Gradio
• Docker

---

📁 Project Structure

. ├── app.py # Gradio UI + Audio generation ├── blog_summarizer.py # CrewAI pipeline + Ollama LLM ├── agents.py # CrewAI Agents (optional) ├── tasks.py # CrewAI Tasks (optional) ├── podcast_transitions.py # Optional podcast formatting ├── styles.css # Custom UI styling ├── requirements.txt ├── Dockerfile ├── .env # API keys (NOT pushed to GitHub) └── README.md

---

🔐 Environment Variables (.env)

Create a .env file in the project root:

FIRECRAWL_API_KEY=your_firecrawl_key
ELEVENLABS_API_KEY=your_elevenlabs_key

⚠️ Never commit .env to GitHub.
🐳 Running with Docker (Local)
1️⃣ Start Ollama Container

sudo docker run -d -p 11434:11434 --name ollama ollama/ollama

Pull the model:

sudo docker exec -it ollama ollama pull deepseek-r1:1.5b

2️⃣ Build Your App Image

sudo docker build -t ai-podcast .

3️⃣ Run the App

sudo docker run -it -p 7777:7777 --env-file .env ai-podcast

Open in browser:

http://localhost:7777

🌐 How It Works

    User pastes a blog URL

    CrewAI agent scrapes content

    Second agent summarizes it into podcast script

    ElevenLabs converts text to MP3 voice

    Gradio returns summary + playable audio

🧪 Sample Test Blog

You can test with:

https://martinfowler.com/articles/microservices.html

🧩 Requirements

Installed automatically via Docker:

crewai
crewai-tools
langchain-ollama
firecrawl
gradio
python-dotenv
requests
litellm
elevenlabs

⚠️ Limitations

    Requires active internet for Firecrawl + ElevenLabs

    Ollama model must be pulled before first run

    Free ElevenLabs quota is limited

📌 Future Improvements

    ✅ Multiple voice support

    ✅ Podcast background music

    ✅ Episode auto-formatting

    ✅ Cloud deployment

    ✅ YouTube auto-publishing

👤 Author

Samprit Roy
GitHub: https://github.com/Samprit74
⭐ Support

If you like this project:

    ⭐ Star the repository

    🧑‍💻 Fork it

    📢 Share it with others

Happy Building! 🚀