Case Study

A Fully Autonomous AI-Powered Podcast

28+ episodes produced with zero manual intervention — from script generation to multi-platform publishing.

28+

Episodes

0

Manual Steps

10+

Platforms

~10min

Per Episode

The Challenge

Can AI run an entire podcast?

Creating a podcast typically requires hours of scripting, recording, editing, and publishing per episode. I wanted to prove that AI could handle the entire pipeline — every step, autonomously, at broadcast quality.

The Pipeline

Six stages. Zero human input.

01

Script Generation

LLM creates unique callers with personalities, topics, and dialogue via OpenRouter.

02

Voice Synthesis

ElevenLabs TTS generates distinct voices for each caller character.

03

Audio Production

REAPER DAW with custom Lua scripts — automated silence removal, normalization, music bed trimming, ad/ident insertion, crossfades.

04

Publishing

Castopod (self-hosted) distributes to Spotify, Apple Podcasts, YouTube, RSS.

05

Social

Postiz (self-hosted) auto-publishes to 10+ social networks.

06

Monitoring

Uptime Kuma tracks every service; custom stats dashboard on CDN.

Tech Stack

Built on open source and APIs.

Python OpenRouter ElevenLabs REAPER + Lua Castopod Postiz Cloudflare CDN Uptime Kuma

The Results

Fully autonomous content production.

28+ episodes produced autonomously

Each episode takes ~10 minutes of compute

Entire pipeline runs on self-hosted infrastructure

Zero recurring platform fees beyond API costs

Want to automate your content pipeline?

Let's talk about what AI automation can do for your content production.

Get in Touch