Menu
Book a Call
← All work

One Podcast Episode In, a Content Engine Out

A health podcast was losing hours per episode to manual transcription, blogging, and clip-cutting. I built the pipeline that automates the whole chain, with a 12-tab admin UI the owner controls.

Status

Client Delivery

Domain

AI Automation

Headline result

Episode-to-blog-to-clips pipeline live; 6 episodes processed; phases 1–5 of 8 delivered

Demonstrates

AI content automation Admin-first product design Teach-to-maintain handoff

Representative stack

WordPress Whisper GPT-4.1 ffmpeg PHP

Input

  • Episode video acquisition

Pipeline

  • Whisper transcription
  • GPT-4.1 blog + topic extraction
  • ffmpeg clip factory with captions

Output

  • WordPress drafts
  • Captioned social clips
  • 12-tab admin UI
From raw episode video to publishable blog and captioned clips, owner-operated

Situation

A health podcast brand run by a non-technical solo founder. Every episode, 22 to 46 minutes of recorded conversation, generated hours of downstream manual work: transcription, blog writing, finding quotable moments, cutting and captioning clips for social.

Problem

The bottleneck was not making the podcast; it was everything after. Each episode’s post-production content work consumed enough hours that output lagged recording, and the backlog of unexploited episodes kept growing. Classic owner-led content economics: the asset exists, the leverage doesn’t.

Approach

Automate the chain end to end as a WordPress plugin living where the owner already works. Episodes flow through Whisper transcription, GPT-4.1 generates blog drafts and topic extractions, and an ffmpeg-based clip factory cuts and captions short-form video. Every stage is operable from a 12-tab admin interface. The owner runs the machine, not me.

Architecture and key decisions

  • Build inside the owner’s existing tool. A WordPress plugin beats a separate SaaS the founder would have to learn, log into, and pay for forever.
  • Admin UI for every stage. All 12 tabs exist so a non-technical owner can configure, run, review, and fix the pipeline without a developer. The system is handed over, not rented out.
  • Human review between generation and publish. Drafts and clips land for approval; the pipeline accelerates the owner’s judgment rather than replacing it.
  • Honest hardening status. Phases 1 through 5 of 8 are built; the pipeline is real but not yet battle-tested at volume, and the client knows exactly which is which.

What shipped

The WordPress automation plugin (video acquisition → transcription → blog and topics → clip factory with captions), the 12-tab admin interface, and a client-facing content strategy deck.

Outcome

Six episodes processed through the live pipeline. The before/after hours-per-episode measurement is being captured on upcoming episodes. I publish measured time savings, not estimates.

What this demonstrates

The owner-led automation pattern I deliver repeatedly: find the manual chain that caps output, automate it inside tools the client already owns, put an admin surface on every moving part, and hand over a system the owner can run and adjust without me.