Content

Podcast Content Repurposing with AI Agents

Hand a podcast episode to an AI agent: auto-transcribe the audio, summarise and slice it into written content, then voice short audio clips — one episode becomes many formats, no manual editing.

Podcast Content Repurposing with AI Agents

Tool names in this guide are for reference only and may have changed or been renamed. Please use the latest tools in the tool marketplace when implementing.

Key takeaways

  • Chain transcription, summary and voicing so one episode becomes many formats
  • Turn a long recording into written notes and short audio in natural language
  • Once the flow is stable, batch the same instruction across a whole season

Tools used

Why one episode is so hard to fully mine

When an episode is done, most of its value stays locked up. The audio sits on a platform, listeners hear it once, and that's it. Turning it into a blog post, a social audio clip, or show notes means starting over for each format: transcribe it by hand or pay for a service, summarise section by section, pull the best quotes, then record a fresh short clip as a teaser. Spinning one episode into three or four formats means running the whole production again three or four times — all the time goes into repetition.

The real pain isn't that you can't produce the formats — it's that every format starts from scratch. Chaining transcription, summarisation and re-voicing into one automated pipeline lets an AI agent connected to these capabilities produce many formats from a single episode. You just say what you want in natural language.

The workflow in four steps

  1. Hand over the audio: give the agent the podcast audio file.
  2. Transcribe: use a transcription tool to turn the audio into a timed text transcript.
  3. Summarise and slice: have the agent summarise, pull quotes and cut the transcript into publishable pieces.
  4. Re-voice into formats: use a speech tool to voice the summary into short audio, shipped alongside the written output.

How to do each step

Step 1: Give the audio to the agent

Prepare the episode's audio file (or a reachable audio link) and say in natural language which formats you want — a written summary, a short audio teaser, or both.

use ae to repurpose this podcast episode at example.com/ep.mp3 into a text summary and a short audio teaser

Step 2: Transcribe the audio

The agent calls a transcription tool such as ae_whisper to turn the audio into text. Transcription is the foundation of the whole chain: with an accurate transcript, everything downstream — summary, slicing — has something to work from. Long audio is processed in segments and returned as a full timed transcript.

use ae to transcribe the audio with ae_whisper

Step 3: Summarise and slice

With the transcript in hand, have the agent summarise it, distil the key points, pull standalone quotes, and cut it into pieces sized for different channels — a one-paragraph show description, a few social captions, a timestamped chapter list. This step is pure text processing, and your instruction decides the output shapes.

use ae to summarise the transcript into show notes and three social captions

Step 4: Re-voice into multiple formats

Hand the summary or teaser script to a speech-synthesis tool such as ae_f5_tts to voice it into a short audio clip — a ready-to-ship audio teaser. At this point one episode has fanned out into a full transcript, written notes, social quotes and a short audio teaser: many formats from the same source audio, produced in parallel.

use ae to voice the teaser script with ae_f5_tts

Manual repurposing vs automated agent fan-out:

Manual repurposing Automated agent fan-out
Transcribe by hand or outsource each episode One line calls ae_whisper to auto-transcribe
Summarise and pick quotes manually Summary + slices in one pass over the transcript
Record a separate short teaser ae_f5_tts voices the teaser directly
Build each format separately One instruction chain produces many formats

How to get started

  1. Sign up — new users get 1,000 free credits.
  2. Connect in-app from an agent like WorkBuddy or Trae by searching for "AgentEarth", or paste a one-line MCP config into Claude Code, Cursor or VSCode.
  3. Give the agent the podcast audio, say which formats you want in natural language, and let it run from transcription through to voicing.

FAQ

Can it slice automatically? Yes. Once the transcript exists, state in your instruction which pieces you want (show description, social captions, chapter list) and the agent summarises and cuts the text into those formats.

Can it use different voices? Yes. The speech tool supports different voices — name the style you want in your instruction, then listen back to confirm it matches.

Does it support long audio? Yes. Long audio is processed in segments and stitched into a full transcript. Run the whole chain on one episode first, then apply the same instruction across a whole season in batch.

Browse the tools · Access guide