How to build an AI content repurposing engine for video

The problem it solves

Marketing teams invest in podcasts, webinars and video, then publish them once. Repurposing is the highest-leverage content work, but copywriters spend hours rewatching and transcribing. Agencies want a repeatable pipeline that produces quality first drafts across channels for every client.

What to include in your first version

• ■Blog article, newsletter, LinkedIn and X posts from one video
• ■Short-form clip suggestions with timestamps and hooks
• ■Brand voice profiles per client
• ■An approval queue with edit history
• ■Export to scheduling tools and CMSs

How to build it, step by step

1. 01

   Fetch the transcript

   Call /transcribe with format.paragraphs and format.timestamp. Use /batch when a client sends a backlog.

2. 02

   Extract the key ideas

   Ask an LLM to list the strongest ideas, quotes and stories with their timestamps. Every format starts from this list.

3. 03

   Generate each format

   Run a dedicated prompt per format, each with the client's voice guide and examples of past posts.

4. 04

   Review and schedule

   Show drafts in an approval queue, then export approved content to the client's scheduler or CMS.

Starter code

This idea is built on the "Recipe 1: One video in, structured output out" pattern. Swap in your own channels, prompts and storage.

import requests

API = "https://www.youtubetranscript.dev/api/v2"
HEADERS = {"Authorization": "Bearer yt_sk_live_YOUR_KEY"}

res = requests.post(f"{API}/transcribe", headers=HEADERS, json={
    "video": "https://www.youtube.com/watch?v=VIDEO_ID",
    "format": {"timestamp": True, "paragraphs": True},
}).json()

transcript = res["data"]["transcript"]
print(transcript["text"][:500])        # plain text for your LLM prompt
for seg in transcript["segments"][:3]:  # timestamps for deep links
    print(seg["start"], seg["text"])

How it makes money

Sell seat-based plans to agencies and content teams, or run it as a done-for-you service where software lets a small team deliver for many clients. Pricing by videos processed per month aligns with your costs.

Mistakes to avoid

• ✕Generating every format from the raw transcript: extract key ideas first for more focused output.
• ✕Ignoring brand voice: generic AI copy is the main reason teams stop using repurposing tools.
• ✕Skipping human review: keep an approval step, especially for client work.

Questions about this idea