Adds broadcast-standard timecodes to plain transcripts at regular intervals or speaker changes for documentary editing and compliance delivery.
How this skill is triggered — by the user, by Claude, or both
Slash command
/autopunk-media-skills:transcript-timecoderThe summary Claude sees in its skill listing — used to decide when to auto-load this skill
Adds broadcast-standard timecodes to a plain transcript at regular intervals or at speaker changes, producing a timecoded transcript formatted for use in documentary editing, paper edits, and compliance delivery.
Adds broadcast-standard timecodes to a plain transcript at regular intervals or at speaker changes, producing a timecoded transcript formatted for use in documentary editing, paper edits, and compliance delivery.
Required:
Optional:
Analyses the transcript structure. Identifies speaker changes, paragraph breaks, natural pauses (indicated by ellipses, dashes, or "[pause]" markers), and approximate word count per section. Uses average speech rate (~150 words per minute for interview speech, ~130 for considered/emotional speech, ~170 for fast conversational speech) to estimate elapsed time from the starting timecode.
Inserts timecodes at the specified intervals. If the user requests speaker-change timecodes, places a timecode at every speaker transition. If fixed-interval timecodes are requested, inserts them at the specified time gap (e.g., every 30 seconds), splitting the transcript text at the nearest sentence boundary to avoid mid-word breaks. If both are requested, combines them — speaker changes always get a timecode, and fixed-interval timecodes are added within long single-speaker passages.
Formats timecodes to broadcast standard. Uses HH:MM:SS:FF format (where FF = frames) at the specified frame rate. For 25fps: frames run 00–24. For 29.97fps drop-frame: uses semicolons (HH:MM:SS;FF) as per SMPTE convention. Timecodes are placed on their own line above the corresponding text, or inline at the start of the text block, depending on the user's formatting preference (defaults to own-line placement).
Estimates rather than fabricates. Every timecode is clearly marked as an estimate based on speech-rate calculation from the transcript text. The skill does not have access to the actual audio/video file — it cannot provide frame-accurate timecodes. It provides working references that are typically accurate to within 5–15 seconds per 10 minutes of material, which is sufficient for paper edits and rough logging but not for final conforming.
Outputs a clean, production-ready document. The final transcript is formatted for immediate use: clear speaker labels, consistent timecode placement, no extraneous formatting. Includes a header noting the source, start timecode, frame rate, and a disclaimer that timecodes are estimated.
Timecoded transcript document. Header block with: source description, start TC, frame rate, and estimation disclaimer. Body: timecodes on their own line (bold or bracketed) followed by speaker label and text. Speaker labels are consistent throughout (UPPERCASE or as provided by user). Paragraphs break at speaker changes and at timecode intervals. No line exceeds a reasonable reading width. For a 30-minute interview, expect approximately 30–60 timecode markers depending on interval setting.
TIMECODED TRANSCRIPT
Source: [Description]
Start TC: [HH:MM:SS:FF] | Frame rate: [25fps]
Note: Timecodes are estimated from speech rate and transcript text.
Accuracy: ±5–15 seconds per 10 minutes. Verify against source media.
---
[HH:MM:SS:FF]
SPEAKER NAME:
Transcript text for this section...
[HH:MM:SS:FF]
SPEAKER NAME:
Transcript text continues...
Transcript:
Maria Castellano: I grew up on this river. My grandfather fished here, my mother swam here every summer, and I learned to row here when I was six. It was the centre of everything.
Interviewer: And when did that start to change?
Maria Castellano: The factory opened in eighty-nine. At first nobody thought much of it. Jobs, you know? Everyone was glad. But within a couple of years the fish started dying. You could smell it. The water turned this sort of grey-green colour and my mother said, don't you dare go near that river. And we didn't. Nobody did. A whole generation of kids in this town grew up without ever touching the water.
Interviewer: Has anything improved since the cleanup programme began?
Maria Castellano: On paper, yes. The readings are better. They tell us it's safe now. But trust doesn't come back with a government report. I still won't eat a fish from that river. My daughter has never swum in it. Maybe her children will, I don't know. But for my generation — that river is gone. Even though it's still there.
Start timecode: 01:03:22:00 Frame rate: 25fps Interval: At every speaker change
TIMECODED TRANSCRIPT Source: Interview — Maria Castellano Start TC: 01:03:22:00 | Frame rate: 25fps Note: Timecodes are estimated from speech rate and transcript text. Accuracy: ±5–15 seconds per 10 minutes. Verify against source media.
[01:03:22:00] MARIA CASTELLANO: I grew up on this river. My grandfather fished here, my mother swam here every summer, and I learned to row here when I was six. It was the centre of everything.
[01:03:36:00] INTERVIEWER: And when did that start to change?
[01:03:39:00] MARIA CASTELLANO: The factory opened in eighty-nine. At first nobody thought much of it. Jobs, you know? Everyone was glad. But within a couple of years the fish started dying. You could smell it. The water turned this sort of grey-green colour and my mother said, don't you dare go near that river. And we didn't. Nobody did. A whole generation of kids in this town grew up without ever touching the water.
[01:04:24:00] INTERVIEWER: Has anything improved since the cleanup programme began?
[01:04:28:00] MARIA CASTELLANO: On paper, yes. The readings are better. They tell us it's safe now. But trust doesn't come back with a government report. I still won't eat a fish from that river. My daughter has never swum in it. Maybe her children will, I don't know. But for my generation — that river is gone. Even though it's still there.
[01:04:58:00] — END OF EXCERPT —
npx claudepluginhub ur-grue/autopunk-media-skills --plugin autopunk-media-skillsTransforms research notes, transcripts, or articles into broadcast-ready voice-over scripts with short sentences, active voice, and timing estimates for picture.
Converts raw meeting transcript .txt files into structured .md notes with metadata, TL;DR, key topics, action items, and quotes. Useful for processing meeting recordings or chat logs.
Convert between 30+ caption/subtitle formats (SRT, VTT, ASS, JSON, TextGrid, LRC, FCPXML, Premiere, …) and shift timing. Trigger on "convert captions", "SRT to VTT", "转换字幕格式", "shift timing", "ASS styling", "karaoke effect", "导入Premiere", or any caption-format question. Do NOT trigger to fix timing accuracy (`/lai-align`) or translate (`/lai-translate`).