If you edit a podcast, you know the grind: hours of raw audio, one usable nugget, and a mountain of show notes that nobody has time to write. These three prompts hand the busywork to AI so you keep the good parts: the voice, the story, the moment.
[Claude]
You are a podcast clip finder. Given the full transcript of my latest episode (paste below), identify the 5 strongest 30-to-60-second moments: emotional highs, surprising stats, and quotable lines. For each, give the exact start and end timestamps, a one-line hook title, and a 140-character social caption. Output as a table. Then pick the single best clip and suggest a 15-second cold-open that hooks a listener who has never heard the show.
[ChatGPT]
Act as my podcast producer. Below is the messy transcript of my new episode. Produce a clean set of show notes: a 3-sentence summary, 5 bullet takeaways in listener language, 3 timestamped chapter markers with titles, and an outro paragraph that teases next week. Skip fluff and tangents. Keep every bullet under 15 words. Format as plain markdown I can paste straight into my hosting platform.
[Cursor]
Build me a lightweight Python script that reads a transcript .txt file and generates a promotional thread for X. It should split the text into short sentences, score each for quote-worthiness (caps, questions, numbers), and output the top 8 lines as a numbered thread with a bare-domain footer reading aitoolalliance.com. Write the script to a single file, add clear comments, and show me the exact CLI command to run it.
Full archive at aitoolalliance.com/tag/prompt-pack.