I Built A Python SaaS with AI

Episode Deep Dive

Guest

Fred Tubiermont is an entrepreneur with a diverse career, spanning music festival production, radio DJing, startup leadership, and even running a circus. At nearly 50, he dove into Python development after decades of working alongside programmers. In under a year of hands-on coding (with heavy assistance from AI tools), Fred built an entire SaaS product—AI Jingle Maker—which now serves paying customers who need quick, professional, AI-generated audio branding.

1. Fred’s Transition into Python

• Late Start, Quick Progress: Fred began learning Python at age 48, moving from local automation scripts to full web apps in just a few months.
• AI as an Enabler: Fred credits tools like ChatGPT for bridging his coding knowledge gaps and accelerating his path to a running SaaS.

2. Building the AI Jingle Maker

• Purpose: Create custom sweepers, radio station IDs, podcast intros, and audio promos quickly by blending AI voiceovers, background music, and special audio effects.
• Tech Stack:
  • Flask for the web framework (flask.palletsprojects.com)
  • Vanilla JavaScript + Tailwind CSS for front-end (tailwindcss.com)
  • Railway for deployment (railway.app)
  • Text-to-speech APIs (Fred integrates existing text-to-speech models rather than building his own)
• Business Model: The product is monetized through credits for each jingle or audio promo generated.

3. Using AI to Code

• ChatGPT (chat.openai.com): Fred frequently asked it for boilerplate Flask examples, step-by-step instructions, and layout suggestions for HTML, CSS, and JavaScript.
• GitHub Copilot (github.com/features/copilot): Used inside his IDE (mainly PyCharm) to auto-complete and refactor code quickly, especially repetitive tasks.

4. Simplifying Deployment with Railway

• Single-Platform Approach: Fred hosts both his Flask app and database on Railway.app, keeping things straightforward (similar to Heroku but with integrated database services).
• Low Costs: Even with an app serving hundreds of daily users, his monthly spend has remained under $20, making it a compelling alternative to bigger platforms.

5. No-Code / Low-Code vs. Real Code

• Initial Attraction: Fred once turned to no-code platforms (Bubble, Zapier/Make) but found them limiting on pricing, performance, and customization.
• Full Control: By using Python and writing custom code (often generated by ChatGPT), Fred can address new features in hours rather than days, with total UI flexibility.

6. Additional Projects and Channels

• Call Me Fred YouTube Channel: Fred shares tutorials on coding with AI, building Chrome extensions, and behind-the-scenes looks at AI Jingle Maker.
  • YouTube Channel
• AI Coding Club Newsletter: Fred’s upcoming newsletter focusing on techniques for integrating AI into Python coding.
• AI Show Maker (Coming Soon): A sister application to AI Jingle Maker, designed for crafting full-length audio segments or podcast episodes in a modular way.

Relevant Links

• AI Jingle Maker: https://aijinglemaker.com
• Flask: https://flask.palletsprojects.com/
• Railway: https://railway.app
• Tailwind CSS: https://tailwindcss.com/
• ChatGPT: https://chat.openai.com
• GitHub Copilot: https://github.com/features/copilot
• Call Me Fred (YouTube): https://www.youtube.com/@callmefred

Overall Takeaway

Fred’s story illustrates how AI-powered coding assistance can radically shorten the learning curve, even for those starting software development later in life. By combining Python, frameworks like Flask, and deployment services such as Railway, it’s now possible for solo developers—seasoned or brand-new—to deliver polished, revenue-generating SaaS products in record time.