I Was Drowning in Podcast Admin, So I Built an Edge AI Tool

Description and use

PodCrisp, built on Cloudflare Workers and Llama‑3, instantly generates titles, show notes, and SEO keywords from a podcast transcript, cutting manual admin from hours to minutes.

Problem

Podcast creators spent up to three hours per episode on manual admin tasks such as title creation, transcript summarization, and SEO‑friendly descriptions, reducing time for content creation.

Outcome

The PodCrisp tool reduced post‑production admin from hours to minutes, providing instant, structured metadata and improving podcast discoverability.

Workflow

• 1. Record episode and produce a transcript.
• 2. Send transcript to Cloudflare Worker endpoint.
• 3. Worker calls Llama‑3‑8b‑instruct via Cloudflare AI with a structured system prompt to generate show notes, title options, and SEO keywords.
• 4. Receive JSON response with formatted metadata.
• 5. Publish metadata to podcast hosting platform.

Evidence

I was spending upwards of three hours after every single episode just doing admin work.

Source

I wanted a unified pipeline that could take a raw transcript and instantly output structured, high‑quality metadata.

Source

I built the entire stack on Cloudflare Workers, leveraging their AI inference capabilities at the edge.

Source

Using Llama 3 at the edge via Cloudflare was a practical choice. It’s fast enough to handle the 800‑token limit for show notes without timing out.

Source