Making Waves: We build a podcast studio to accelerate learning

BD
Bobby Dean
May 11, 2026

We had a problem and fixed it. Read this use case to discover the speed of AI delivery.

Making Waves: We build a podcast studio to accelerate learning

I Built a Podcast Studio Because I Needed an Audio Study Guide

Ninety minutes. One way.

That's my daily commute — and far from wasted time. I'd been using Grok's text-to-speech app to have real conversations about tech questions on the road. It became a habit. The voice interaction made complex topics click in a way that reading never quite did at a desk.

Then I signed up for the Claude Certified Architect — Foundations exam.

The material was dense. Five exam domains. Six scenario archetypes. Agentic patterns, retrieval, evaluation, safety, deployment. The kind of content that rewards repetition. And sitting with Grok every morning gave me an idea — what if I took all of the study material for this new exam and converted it into a proper audio guide I could listen to on the drive?

I went looking for one first. Blog posts existed. Docs existed. YouTube videos that assumed I was sitting in front of a screen. But no one had cut a clean, episodic, listen-while-you-commute audio version of the curriculum. Fair enough — the certification is new. The community hasn't had time.

So I decided to make one for myself.

The Hack That Wasn't Supposed to Become a Product

The MVP I had in mind was almost embarrassingly small: an RSS feed I'd subscribe to from my own phone. Draft each episode as a script, run it through TTS, drop the MP3 in S3, hand-write the XML. Six episodes, two weekends, a study tool for an audience of one.

That plan lasted about a week.

Once I started writing the episodes, the holes in my hand-rolled pipeline got expensive fast. I wanted consistent presenter voices across episodes. I wanted scripts grounded in a series bible so they didn't drift. I wanted to regenerate just the audio when I tweaked a single paragraph instead of re-rendering everything. I used Eleven Labs Text to Speech, but after I maxed out my usage, I switched to XAI (since I was already talking to their voice "Ara" with their Grok app)

Then I wanted more... (this vibe coding thing can be addictive!) I wanted proper bookends — intro, outro, a beat of silence between them and the body. I wanted to publish to Spotify and Apple Podcasts without hand-editing iTunes-flavored RSS for the fifth time.

What I really wanted was a podcast studio.

So I built one.

The Loop That Got Recursive

Here's where it got interesting: the certification I was studying for is about how to architect production systems with Claude. The thing I was building to help me study was itself one of those systems — and every architectural decision I made was an answer to a question I was about to be tested on.

Chained agent pipelines with a planner, a writer, a formatter? That's exam Domain 1. Retrieval-augmented generation over my own source documents? Domain 3. Evaluation gates and human-in-the-loop approvals before audio assembly burns tokens? Domains 4 and 5.

I stopped studying for the exam. I was taking it, live, against a real codebase, every day.

What WaveMaker Is Now

WaveMaker is an AI-native podcast studio. It takes you from a series concept to a published show on Spotify and Apple Podcasts without leaving the browser:

  • Plan the series with an AI copilot that drafts the bible, presenters, voice profiles, and a starting set of episode stubs
  • Write each episode with a chained agent pipeline — outline, script, TTS-ready dialogue — grounded in your own source documents
  • Voice the show with consistent, multi-presenter audio assembly, branded intros and outros baked in at publish time
  • Ship to Spotify, Apple Podcasts, and Amazon Music with a single RSS feed that's already iTunes-compliant

This is what AI-native actually means — not AI bolted onto an existing workflow, but a system where the intelligence is load-bearing from the first line of architecture.

Listen to the Show That Built the Studio

The series is called NoLimitz Podcast — Preparing for the Certified Claude Architect Exam. Eight episodes are live right now: an origin episode, then one per exam domain with the scenario walkthroughs woven in. Casual, honest, and built for people who think in systems.

If you're a technical architect with a commute and an exam date — give it a listen:

Start with episode one. The rest is best heard in order.

If WaveMaker sounds like something your team needs — internal training, customer enablement, an ongoing learning podcast for a domain that doesn't have one yet — reach out. I'll be launching it soon, but if you want early access, let me know.

Bobby Dean

  • bobby@nolimitz.io

Building an AI assistant, agent, or chatbot for your business?

We help you get it from prototype to production — with the observability and governance layer built in from the start.

Contact Us