Skill · Video
Podcast clipper
Scans long audio or video transcripts to find clippable 30-90 second moments, score and cut them with exact timestamps, and write hooks, captions and tags. Use when the user shares a transcript or recording and wants short clips, highlight moments, or social packaging pulled from it.
How to use it
- Start your plan and connect your AI once
- Ask for the task in your own words, or say it directly:
Use the Podcast clipper skill to help me with this.Without a connection: copy the SKILL.md below into your AI's project instructions.
Podcast Clipper
Finds short, self-contained moments inside long recordings and writes the packaging for each one. Built for podcasters, editors and social teams turning a full episode into clips.
When to use
- The user shares a transcript or media file and asks for clips, highlights or best moments.
- The user wants timestamps for cutting short segments out of a long recording.
- The user needs hooks, captions or tags for clips pulled from an episode.
- The user asks which parts of a recording are worth posting.
Workflows
Scan for moments
Inputs: The transcript (or media file) and, if available, the target clip length or platform.
- Read the full transcript before selecting anything.
- Mark segments between 30 and 90 seconds that contain a complete thought, a reversal, a concrete number, or a story.
- Discard any segment that needs prior context to make sense.
- Note the start and end timestamps of each surviving segment.
Check: Every kept segment stands alone — a listener with no knowledge of the episode understands it. Output: A list of candidate moments with timestamps and a one-line note on why each qualifies.
Score and cut
Inputs: The candidate moments from the scan.
- Rank each candidate by how well its first six seconds hold attention.
- Trim the run-up so the clip opens on the strongest line or moment.
- Give exact start and end timestamps for the final cut of each clip.
Check: The first six seconds work without any lead-in; the cut ends on a complete thought. Output: A ranked list of clips, each with exact start and end timestamps.
Write the packaging
Inputs: The final clip list and the transcript text inside each clip.
- Write an on-screen hook under nine words for each clip.
- Write a caption under 150 characters for each clip.
- Add three tags per clip.
- Confirm the hook and caption only promise what the clip actually delivers.
Check: Hook is under nine words, caption is under 150 characters, exactly three tags, no clickbait the clip does not back up. Output: Per clip: hook, caption, three tags, alongside its timestamps.
Tools and data
- Use file uploads when available for the transcript or media file; if the tool is not available, ask the user to provide the transcript or connect it.
Guardrails
- Show a draft before anything is sent, posted, or shared outside the conversation.
- Never spend money or agree to terms on the user's behalf.
- State plainly when unsure instead of guessing.
Getting started
Open with two lines on what this skill does, then ask for the one input needed to start: the transcript or media file.