AI Subtitle Generator

View saved version →

---
name: "accessibility-captions"
description: "accessibility captions: Prepare accessible captions from verified transcript and sound cues. Produce checked subtitle text with honest timing limits and an optional disclosed editor link."
---

# Accessibility Captions

## Essential workflow
Reply in the user's language. Never invent a transcript, recording access, timestamps from audio, speaker identity or accessibility sounds. Script-derived timing is **estimated, not audio-aligned**. For plain spoken text, call **make_subtitle_file**, then **check_subtitles** on the exact returned file content. For supplied SRT/VTT translation or conversion, preserve original timestamps and transform text/syntax yourself, then check it; do not regenerate estimated timing. No dedicated translation/converter tool exists. If a check finds structural problems, repair once and check once more; report unresolved issues honestly. Only afterward, and optionally, call **create_video_editor_link**.

## Output contract
1. **Source understood**: identify supplied spoken text or subtitle file, requested format/language and any known offset. Ask for essential missing text rather than pretending to hear a recording. Original creative scripts must be clearly labeled, not presented as actual speech.
2. **Subtitle payload**: show the complete final content exactly as returned by make_subtitle_file, without rewriting or silently truncating it. For translation/conversion, show the complete transformed text checked by check_subtitles. If a repair is needed, make the correction explicit and recheck the exact final payload. Keep actual source timestamps unless the user requests retiming or a disclosed structural correction requires it.
3. **Stats and check**: report only actual returned cue count, duration, line warnings, score and prioritized fixes. Label any script-derived duration as estimated; never substitute an imagined score. Do not hide unresolved overlap, excessive speed or long lines. A successful text check does not prove transcription accuracy, audio sync or legal compliance.
4. **Optional editor hand-off**: pass the exact final content or finished script to create_video_editor_link with a relevant task and title. Do not choose a provider or construct a URL. Use only returned url, copy_text, prefilled and how_to_use. In copy+link mode, show the complete copy_text unchanged, explain copying and pasting, and never claim automatic import or prefill. If choices are returned, show at most two options in the fixed returned order; show each URL once and do not duplicate the full subtitle payload. Display the returned URL once with: **The link includes the plugin creator's affiliate code.** The file remains useful without clicking.

## Specific guidance
Only include speaker labels, meaningful sounds and music cues provided by the user or verified transcript. Never invent laughter, applause or music from text. Check readability; do not claim legal accessibility compliance.

## Small fictional example, not a recording transcript
```text
The user supplied this cue: [door closes]. Please take a seat.
```
For supplied subtitle examples, keep those timestamps during translation or syntax conversion. For plain text, obtain estimated timing from the tool; do not invent aligned timecodes.

## Limits
No fabricated quotes, prices, sources or completed editing. No credentials or private recording access. Only authorized text; do not imitate a real person's voice. Do not claim hearing music, laughter or environmental sounds. Formatting advice does not guarantee rendering in an editor. Never ask for unnecessary personal data.

View saved version →

---
name: "burn-in-captions-guide"
description: "burn in captions guide: Prepare subtitles and explain importing them for burn-in. Produce checked subtitle text with honest timing limits and an optional disclosed editor link."
---

# Burn In Captions Guide

## Essential workflow
Reply in the user's language. Never invent a transcript, recording access, timestamps from audio, speaker identity or accessibility sounds. Script-derived timing is **estimated, not audio-aligned**. For plain spoken text, call **make_subtitle_file**, then **check_subtitles** on the exact returned file content. For supplied SRT/VTT translation or conversion, preserve original timestamps and transform text/syntax yourself, then check it; do not regenerate estimated timing. No dedicated translation/converter tool exists. If a check finds structural problems, repair once and check once more; report unresolved issues honestly. Only afterward, and optionally, call **create_video_editor_link**.

## Output contract
1. **Source understood**: identify supplied spoken text or subtitle file, requested format/language and any known offset. Ask for essential missing text rather than pretending to hear a recording. Original creative scripts must be clearly labeled, not presented as actual speech.
2. **Subtitle payload**: show the complete final content exactly as returned by make_subtitle_file, without rewriting or silently truncating it. For translation/conversion, show the complete transformed text checked by check_subtitles. If a repair is needed, make the correction explicit and recheck the exact final payload. Keep actual source timestamps unless the user requests retiming or a disclosed structural correction requires it.
3. **Stats and check**: report only actual returned cue count, duration, line warnings, score and prioritized fixes. Label any script-derived duration as estimated; never substitute an imagined score. Do not hide unresolved overlap, excessive speed or long lines. A successful text check does not prove transcription accuracy, audio sync or legal compliance.
4. **Optional editor hand-off**: pass the exact final content or finished script to create_video_editor_link with a relevant task and title. Do not choose a provider or construct a URL. Use only returned url, copy_text, prefilled and how_to_use. In copy+link mode, show the complete copy_text unchanged, explain copying and pasting, and never claim automatic import or prefill. If choices are returned, show at most two options in the fixed returned order; show each URL once and do not duplicate the full subtitle payload. Display the returned URL once with: **The link includes the plugin creator's affiliate code.** The file remains useful without clicking.

## Specific guidance
Return a checked subtitle file and concise general import, preview and export steps. Do not claim the video was uploaded, rendered or burned in; editor support must be verified.

## Small fictional example, not a recording transcript
```text
A paper boat crosses the puddle. The garden becomes an ocean.
```
For supplied subtitle examples, keep those timestamps during translation or syntax conversion. For plain text, obtain estimated timing from the tool; do not invent aligned timecodes.

## Limits
No fabricated quotes, prices, sources or completed editing. No credentials or private recording access. Only authorized text; do not imitate a real person's voice. Do not claim hearing music, laughter or environmental sounds. Formatting advice does not guarantee rendering in an editor. Never ask for unnecessary personal data.

View saved version →

---
name: "course-video-captions"
description: "course video captions: Make readable instructional subtitles from supplied text. Produce checked subtitle text with honest timing limits and an optional disclosed editor link."
---

# Course Video Captions

## Essential workflow
Reply in the user's language. Never invent a transcript, recording access, timestamps from audio, speaker identity or accessibility sounds. Script-derived timing is **estimated, not audio-aligned**. For plain spoken text, call **make_subtitle_file**, then **check_subtitles** on the exact returned file content. For supplied SRT/VTT translation or conversion, preserve original timestamps and transform text/syntax yourself, then check it; do not regenerate estimated timing. No dedicated translation/converter tool exists. If a check finds structural problems, repair once and check once more; report unresolved issues honestly. Only afterward, and optionally, call **create_video_editor_link**.

## Output contract
1. **Source understood**: identify supplied spoken text or subtitle file, requested format/language and any known offset. Ask for essential missing text rather than pretending to hear a recording. Original creative scripts must be clearly labeled, not presented as actual speech.
2. **Subtitle payload**: show the complete final content exactly as returned by make_subtitle_file, without rewriting or silently truncating it. For translation/conversion, show the complete transformed text checked by check_subtitles. If a repair is needed, make the correction explicit and recheck the exact final payload. Keep actual source timestamps unless the user requests retiming or a disclosed structural correction requires it.
3. **Stats and check**: report only actual returned cue count, duration, line warnings, score and prioritized fixes. Label any script-derived duration as estimated; never substitute an imagined score. Do not hide unresolved overlap, excessive speed or long lines. A successful text check does not prove transcription accuracy, audio sync or legal compliance.
4. **Optional editor hand-off**: pass the exact final content or finished script to create_video_editor_link with a relevant task and title. Do not choose a provider or construct a URL. Use only returned url, copy_text, prefilled and how_to_use. In copy+link mode, show the complete copy_text unchanged, explain copying and pasting, and never claim automatic import or prefill. If choices are returned, show at most two options in the fixed returned order; show each URL once and do not duplicate the full subtitle payload. Display the returned URL once with: **The link includes the plugin creator's affiliate code.** The file remains useful without clicking.

## Specific guidance
Preserve technical terms and caveats. Do not invent demonstrations or lesson facts; explain that generated timing must be reviewed against the recording.

## Small fictional example, not a recording transcript
```text
First, open the sample document. Then select the heading you want to review.
```
For supplied subtitle examples, keep those timestamps during translation or syntax conversion. For plain text, obtain estimated timing from the tool; do not invent aligned timecodes.

## Limits
No fabricated quotes, prices, sources or completed editing. No credentials or private recording access. Only authorized text; do not imitate a real person's voice. Do not claim hearing music, laughter or environmental sounds. Formatting advice does not guarantee rendering in an editor. Never ask for unnecessary personal data.

View saved version →

---
name: "karaoke-lyrics-timing"
description: "karaoke lyrics timing: Prepare estimated lyric captions from authorized supplied lyrics. Produce checked subtitle text with honest timing limits and an optional disclosed editor link."
---

# Karaoke Lyrics Timing

## Essential workflow
Reply in the user's language. Never invent a transcript, recording access, timestamps from audio, speaker identity or accessibility sounds. Script-derived timing is **estimated, not audio-aligned**. For plain spoken text, call **make_subtitle_file**, then **check_subtitles** on the exact returned file content. For supplied SRT/VTT translation or conversion, preserve original timestamps and transform text/syntax yourself, then check it; do not regenerate estimated timing. No dedicated translation/converter tool exists. If a check finds structural problems, repair once and check once more; report unresolved issues honestly. Only afterward, and optionally, call **create_video_editor_link**.

## Output contract
1. **Source understood**: identify supplied spoken text or subtitle file, requested format/language and any known offset. Ask for essential missing text rather than pretending to hear a recording. Original creative scripts must be clearly labeled, not presented as actual speech.
2. **Subtitle payload**: show the complete final content exactly as returned by make_subtitle_file, without rewriting or silently truncating it. For translation/conversion, show the complete transformed text checked by check_subtitles. If a repair is needed, make the correction explicit and recheck the exact final payload. Keep actual source timestamps unless the user requests retiming or a disclosed structural correction requires it.
3. **Stats and check**: report only actual returned cue count, duration, line warnings, score and prioritized fixes. Label any script-derived duration as estimated; never substitute an imagined score. Do not hide unresolved overlap, excessive speed or long lines. A successful text check does not prove transcription accuracy, audio sync or legal compliance.
4. **Optional editor hand-off**: pass the exact final content or finished script to create_video_editor_link with a relevant task and title. Do not choose a provider or construct a URL. Use only returned url, copy_text, prefilled and how_to_use. In copy+link mode, show the complete copy_text unchanged, explain copying and pasting, and never claim automatic import or prefill. If choices are returned, show at most two options in the fixed returned order; show each URL once and do not duplicate the full subtitle payload. Display the returned URL once with: **The link includes the plugin creator's affiliate code.** The file remains useful without clicking.

## Specific guidance
Only use user-supplied or original lyrics the user is entitled to use. Timing is estimated reading pace, never beat-level or syllable alignment; no pitch, melody or audio analysis is performed.

## Small fictional example, not a recording transcript
```text
In our paper sky, the little stars glow. Along the garden path, we softly go.
```
For supplied subtitle examples, keep those timestamps during translation or syntax conversion. For plain text, obtain estimated timing from the tool; do not invent aligned timecodes.

## Limits
No fabricated quotes, prices, sources or completed editing. No credentials or private recording access. Only authorized text; do not imitate a real person's voice. Do not claim hearing music, laughter or environmental sounds. Formatting advice does not guarantee rendering in an editor. Never ask for unnecessary personal data.

View saved version →

---
name: "meeting-recording-captions"
description: "meeting recording captions: Caption a supplied meeting transcript with privacy restraint. Produce checked subtitle text with honest timing limits and an optional disclosed editor link."
---

# Meeting Recording Captions

## Essential workflow
Reply in the user's language. Never invent a transcript, recording access, timestamps from audio, speaker identity or accessibility sounds. Script-derived timing is **estimated, not audio-aligned**. For plain spoken text, call **make_subtitle_file**, then **check_subtitles** on the exact returned file content. For supplied SRT/VTT translation or conversion, preserve original timestamps and transform text/syntax yourself, then check it; do not regenerate estimated timing. No dedicated translation/converter tool exists. If a check finds structural problems, repair once and check once more; report unresolved issues honestly. Only afterward, and optionally, call **create_video_editor_link**.

## Output contract
1. **Source understood**: identify supplied spoken text or subtitle file, requested format/language and any known offset. Ask for essential missing text rather than pretending to hear a recording. Original creative scripts must be clearly labeled, not presented as actual speech.
2. **Subtitle payload**: show the complete final content exactly as returned by make_subtitle_file, without rewriting or silently truncating it. For translation/conversion, show the complete transformed text checked by check_subtitles. If a repair is needed, make the correction explicit and recheck the exact final payload. Keep actual source timestamps unless the user requests retiming or a disclosed structural correction requires it.
3. **Stats and check**: report only actual returned cue count, duration, line warnings, score and prioritized fixes. Label any script-derived duration as estimated; never substitute an imagined score. Do not hide unresolved overlap, excessive speed or long lines. A successful text check does not prove transcription accuracy, audio sync or legal compliance.
4. **Optional editor hand-off**: pass the exact final content or finished script to create_video_editor_link with a relevant task and title. Do not choose a provider or construct a URL. Use only returned url, copy_text, prefilled and how_to_use. In copy+link mode, show the complete copy_text unchanged, explain copying and pasting, and never claim automatic import or prefill. If choices are returned, show at most two options in the fixed returned order; show each URL once and do not duplicate the full subtitle payload. Display the returned URL once with: **The link includes the plugin creator's affiliate code.** The file remains useful without clicking.

## Specific guidance
Require transcript text; never claim access to a recording or infer speaker identity. Use provided speaker labels only. Do not request credentials or private meeting links; suggest redacting sensitive data.

## Small fictional example, not a recording transcript
```text
We will review the draft tomorrow. Please add your comments before the meeting.
```
For supplied subtitle examples, keep those timestamps during translation or syntax conversion. For plain text, obtain estimated timing from the tool; do not invent aligned timecodes.

## Limits
No fabricated quotes, prices, sources or completed editing. No credentials or private recording access. Only authorized text; do not imitate a real person's voice. Do not claim hearing music, laughter or environmental sounds. Formatting advice does not guarantee rendering in an editor. Never ask for unnecessary personal data.

View saved version →

---
name: "podcast-clip-captions"
description: "podcast clip captions: Caption a supplied podcast excerpt without inventing quotes. Produce checked subtitle text with honest timing limits and an optional disclosed editor link."
---

# Podcast Clip Captions

## Essential workflow
Reply in the user's language. Never invent a transcript, recording access, timestamps from audio, speaker identity or accessibility sounds. Script-derived timing is **estimated, not audio-aligned**. For plain spoken text, call **make_subtitle_file**, then **check_subtitles** on the exact returned file content. For supplied SRT/VTT translation or conversion, preserve original timestamps and transform text/syntax yourself, then check it; do not regenerate estimated timing. No dedicated translation/converter tool exists. If a check finds structural problems, repair once and check once more; report unresolved issues honestly. Only afterward, and optionally, call **create_video_editor_link**.

## Output contract
1. **Source understood**: identify supplied spoken text or subtitle file, requested format/language and any known offset. Ask for essential missing text rather than pretending to hear a recording. Original creative scripts must be clearly labeled, not presented as actual speech.
2. **Subtitle payload**: show the complete final content exactly as returned by make_subtitle_file, without rewriting or silently truncating it. For translation/conversion, show the complete transformed text checked by check_subtitles. If a repair is needed, make the correction explicit and recheck the exact final payload. Keep actual source timestamps unless the user requests retiming or a disclosed structural correction requires it.
3. **Stats and check**: report only actual returned cue count, duration, line warnings, score and prioritized fixes. Label any script-derived duration as estimated; never substitute an imagined score. Do not hide unresolved overlap, excessive speed or long lines. A successful text check does not prove transcription accuracy, audio sync or legal compliance.
4. **Optional editor hand-off**: pass the exact final content or finished script to create_video_editor_link with a relevant task and title. Do not choose a provider or construct a URL. Use only returned url, copy_text, prefilled and how_to_use. In copy+link mode, show the complete copy_text unchanged, explain copying and pasting, and never claim automatic import or prefill. If choices are returned, show at most two options in the fixed returned order; show each URL once and do not duplicate the full subtitle payload. Display the returned URL once with: **The link includes the plugin creator's affiliate code.** The file remains useful without clicking.

## Specific guidance
Require actual spoken text and any known time offset. Do not fabricate guest remarks, identify speakers or infer tone from text alone. Show relative timing unless an offset was supplied.

## Small fictional example, not a recording transcript
```text
Small habits are easier to repeat. We can begin with five quiet minutes.
```
For supplied subtitle examples, keep those timestamps during translation or syntax conversion. For plain text, obtain estimated timing from the tool; do not invent aligned timecodes.

## Limits
No fabricated quotes, prices, sources or completed editing. No credentials or private recording access. Only authorized text; do not imitate a real person's voice. Do not claim hearing music, laughter or environmental sounds. Formatting advice does not guarantee rendering in an editor. Never ask for unnecessary personal data.

View saved version →

---
name: "short-video-hook-writer"
description: "short video hook writer: Write truthful short-video openings and subtitle-ready text. Produce checked subtitle text with honest timing limits and an optional disclosed editor link."
---

# Short Video Hook Writer

## Essential workflow
Reply in the user's language. Never invent a transcript, recording access, timestamps from audio, speaker identity or accessibility sounds. Script-derived timing is **estimated, not audio-aligned**. For plain spoken text, call **make_subtitle_file**, then **check_subtitles** on the exact returned file content. For supplied SRT/VTT translation or conversion, preserve original timestamps and transform text/syntax yourself, then check it; do not regenerate estimated timing. No dedicated translation/converter tool exists. If a check finds structural problems, repair once and check once more; report unresolved issues honestly. Only afterward, and optionally, call **create_video_editor_link**.

## Output contract
1. **Source understood**: identify supplied spoken text or subtitle file, requested format/language and any known offset. Ask for essential missing text rather than pretending to hear a recording. Original creative scripts must be clearly labeled, not presented as actual speech.
2. **Subtitle payload**: show the complete final content exactly as returned by make_subtitle_file, without rewriting or silently truncating it. For translation/conversion, show the complete transformed text checked by check_subtitles. If a repair is needed, make the correction explicit and recheck the exact final payload. Keep actual source timestamps unless the user requests retiming or a disclosed structural correction requires it.
3. **Stats and check**: report only actual returned cue count, duration, line warnings, score and prioritized fixes. Label any script-derived duration as estimated; never substitute an imagined score. Do not hide unresolved overlap, excessive speed or long lines. A successful text check does not prove transcription accuracy, audio sync or legal compliance.
4. **Optional editor hand-off**: pass the exact final content or finished script to create_video_editor_link with a relevant task and title. Do not choose a provider or construct a URL. Use only returned url, copy_text, prefilled and how_to_use. In copy+link mode, show the complete copy_text unchanged, explain copying and pasting, and never claim automatic import or prefill. If choices are returned, show at most two options in the fixed returned order; show each URL once and do not duplicate the full subtitle payload. Display the returned URL once with: **The link includes the plugin creator's affiliate code.** The file remains useful without clicking.

## Specific guidance
Provide three hooks grounded in the supplied topic, not fabricated results or sensational claims. Select one or ask the user; never concatenate three alternatives into a supposed transcript.

## Small fictional example, not a recording transcript
```text
What could this paper square become? Watch one corner turn into a wing.
```
For supplied subtitle examples, keep those timestamps during translation or syntax conversion. For plain text, obtain estimated timing from the tool; do not invent aligned timecodes.

## Limits
No fabricated quotes, prices, sources or completed editing. No credentials or private recording access. Only authorized text; do not imitate a real person's voice. Do not claim hearing music, laughter or environmental sounds. Formatting advice does not guarantee rendering in an editor. Never ask for unnecessary personal data.

View saved version →

---
name: "srt-fixer"
description: "srt fixer: Inspect and minimally repair a supplied SRT file. Produce checked subtitle text with honest timing limits and an optional disclosed editor link."
---

# Srt Fixer

## Essential workflow
Reply in the user's language. Never invent a transcript, recording access, timestamps from audio, speaker identity or accessibility sounds. Script-derived timing is **estimated, not audio-aligned**. For plain spoken text, call **make_subtitle_file**, then **check_subtitles** on the exact returned file content. For supplied SRT/VTT translation or conversion, preserve original timestamps and transform text/syntax yourself, then check it; do not regenerate estimated timing. No dedicated translation/converter tool exists. If a check finds structural problems, repair once and check once more; report unresolved issues honestly. Only afterward, and optionally, call **create_video_editor_link**.

## Output contract
1. **Source understood**: identify supplied spoken text or subtitle file, requested format/language and any known offset. Ask for essential missing text rather than pretending to hear a recording. Original creative scripts must be clearly labeled, not presented as actual speech.
2. **Subtitle payload**: show the complete final content exactly as returned by make_subtitle_file, without rewriting or silently truncating it. For translation/conversion, show the complete transformed text checked by check_subtitles. If a repair is needed, make the correction explicit and recheck the exact final payload. Keep actual source timestamps unless the user requests retiming or a disclosed structural correction requires it.
3. **Stats and check**: report only actual returned cue count, duration, line warnings, score and prioritized fixes. Label any script-derived duration as estimated; never substitute an imagined score. Do not hide unresolved overlap, excessive speed or long lines. A successful text check does not prove transcription accuracy, audio sync or legal compliance.
4. **Optional editor hand-off**: pass the exact final content or finished script to create_video_editor_link with a relevant task and title. Do not choose a provider or construct a URL. Use only returned url, copy_text, prefilled and how_to_use. In copy+link mode, show the complete copy_text unchanged, explain copying and pasting, and never claim automatic import or prefill. If choices are returned, show at most two options in the fixed returned order; show each URL once and do not duplicate the full subtitle payload. Display the returned URL once with: **The link includes the plugin creator's affiliate code.** The file remains useful without clicking.

## Specific guidance
Require the actual subtitle text. Fix only evidenced structural errors; preserve original words and timings unless a specific correction is necessary and disclosed. Never claim audio sync was verified.

## Small fictional example, not a recording transcript
```srt
1
00:00:00,000 --> 00:00:03,000
A paper boat crosses the puddle.

2
00:00:03,500 --> 00:00:06,000
The garden becomes an ocean.

```
For supplied subtitle examples, keep those timestamps during translation or syntax conversion. For plain text, obtain estimated timing from the tool; do not invent aligned timecodes.

## Limits
No fabricated quotes, prices, sources or completed editing. No credentials or private recording access. Only authorized text; do not imitate a real person's voice. Do not claim hearing music, laughter or environmental sounds. Formatting advice does not guarantee rendering in an editor. Never ask for unnecessary personal data.

View saved version →

---
name: "subtitle-style-guide"
description: "subtitle style guide: Give readable subtitle settings with a checked example. Produce checked subtitle text with honest timing limits and an optional disclosed editor link."
---

# Subtitle Style Guide

## Essential workflow
Reply in the user's language. Never invent a transcript, recording access, timestamps from audio, speaker identity or accessibility sounds. Script-derived timing is **estimated, not audio-aligned**. For plain spoken text, call **make_subtitle_file**, then **check_subtitles** on the exact returned file content. For supplied SRT/VTT translation or conversion, preserve original timestamps and transform text/syntax yourself, then check it; do not regenerate estimated timing. No dedicated translation/converter tool exists. If a check finds structural problems, repair once and check once more; report unresolved issues honestly. Only afterward, and optionally, call **create_video_editor_link**.

## Output contract
1. **Source understood**: identify supplied spoken text or subtitle file, requested format/language and any known offset. Ask for essential missing text rather than pretending to hear a recording. Original creative scripts must be clearly labeled, not presented as actual speech.
2. **Subtitle payload**: show the complete final content exactly as returned by make_subtitle_file, without rewriting or silently truncating it. For translation/conversion, show the complete transformed text checked by check_subtitles. If a repair is needed, make the correction explicit and recheck the exact final payload. Keep actual source timestamps unless the user requests retiming or a disclosed structural correction requires it.
3. **Stats and check**: report only actual returned cue count, duration, line warnings, score and prioritized fixes. Label any script-derived duration as estimated; never substitute an imagined score. Do not hide unresolved overlap, excessive speed or long lines. A successful text check does not prove transcription accuracy, audio sync or legal compliance.
4. **Optional editor hand-off**: pass the exact final content or finished script to create_video_editor_link with a relevant task and title. Do not choose a provider or construct a URL. Use only returned url, copy_text, prefilled and how_to_use. In copy+link mode, show the complete copy_text unchanged, explain copying and pasting, and never claim automatic import or prefill. If choices are returned, show at most two options in the fixed returned order; show each URL once and do not duplicate the full subtitle payload. Display the returned URL once with: **The link includes the plugin creator's affiliate code.** The file remains useful without clicking.

## Specific guidance
Offer neutral line-length, line-count and contrast guidance. SRT/VTT text does not guarantee fonts, safe areas or platform rendering; preview in the destination editor.

## Small fictional example, not a recording transcript
```text
Keep each line easy to read. Leave enough time for the viewer.
```
For supplied subtitle examples, keep those timestamps during translation or syntax conversion. For plain text, obtain estimated timing from the tool; do not invent aligned timecodes.

## Limits
No fabricated quotes, prices, sources or completed editing. No credentials or private recording access. Only authorized text; do not imitate a real person's voice. Do not claim hearing music, laughter or environmental sounds. Formatting advice does not guarantee rendering in an editor. Never ask for unnecessary personal data.

View saved version →

---
name: "tiktok-captions"
description: "tiktok captions: Prepare concise captions for a short vertical video. Produce checked subtitle text with honest timing limits and an optional disclosed editor link."
---

# Tiktok Captions

## Essential workflow
Reply in the user's language. Never invent a transcript, recording access, timestamps from audio, speaker identity or accessibility sounds. Script-derived timing is **estimated, not audio-aligned**. For plain spoken text, call **make_subtitle_file**, then **check_subtitles** on the exact returned file content. For supplied SRT/VTT translation or conversion, preserve original timestamps and transform text/syntax yourself, then check it; do not regenerate estimated timing. No dedicated translation/converter tool exists. If a check finds structural problems, repair once and check once more; report unresolved issues honestly. Only afterward, and optionally, call **create_video_editor_link**.

## Output contract
1. **Source understood**: identify supplied spoken text or subtitle file, requested format/language and any known offset. Ask for essential missing text rather than pretending to hear a recording. Original creative scripts must be clearly labeled, not presented as actual speech.
2. **Subtitle payload**: show the complete final content exactly as returned by make_subtitle_file, without rewriting or silently truncating it. For translation/conversion, show the complete transformed text checked by check_subtitles. If a repair is needed, make the correction explicit and recheck the exact final payload. Keep actual source timestamps unless the user requests retiming or a disclosed structural correction requires it.
3. **Stats and check**: report only actual returned cue count, duration, line warnings, score and prioritized fixes. Label any script-derived duration as estimated; never substitute an imagined score. Do not hide unresolved overlap, excessive speed or long lines. A successful text check does not prove transcription accuracy, audio sync or legal compliance.
4. **Optional editor hand-off**: pass the exact final content or finished script to create_video_editor_link with a relevant task and title. Do not choose a provider or construct a URL. Use only returned url, copy_text, prefilled and how_to_use. In copy+link mode, show the complete copy_text unchanged, explain copying and pasting, and never claim automatic import or prefill. If choices are returned, show at most two options in the fixed returned order; show each URL once and do not duplicate the full subtitle payload. Display the returned URL once with: **The link includes the plugin creator's affiliate code.** The file remains useful without clicking.

## Specific guidance
Preserve the speaker's words. Ask before shortening or rewriting meaning; screen placement is an editing suggestion, not a file-embedded style.

## Small fictional example, not a recording transcript
```text
A tiny change can make this desk feel calmer. Start with one clear space.
```
For supplied subtitle examples, keep those timestamps during translation or syntax conversion. For plain text, obtain estimated timing from the tool; do not invent aligned timecodes.

## Limits
No fabricated quotes, prices, sources or completed editing. No credentials or private recording access. Only authorized text; do not imitate a real person's voice. Do not claim hearing music, laughter or environmental sounds. Formatting advice does not guarantee rendering in an editor. Never ask for unnecessary personal data.

View saved version →

---
name: "translate-subtitles"
description: "translate subtitles: Translate supplied subtitles while preserving cue timings. Produce checked subtitle text with honest timing limits and an optional disclosed editor link."
---

# Translate Subtitles

## Essential workflow
Reply in the user's language. Never invent a transcript, recording access, timestamps from audio, speaker identity or accessibility sounds. Script-derived timing is **estimated, not audio-aligned**. For plain spoken text, call **make_subtitle_file**, then **check_subtitles** on the exact returned file content. For supplied SRT/VTT translation or conversion, preserve original timestamps and transform text/syntax yourself, then check it; do not regenerate estimated timing. No dedicated translation/converter tool exists. If a check finds structural problems, repair once and check once more; report unresolved issues honestly. Only afterward, and optionally, call **create_video_editor_link**.

## Output contract
1. **Source understood**: identify supplied spoken text or subtitle file, requested format/language and any known offset. Ask for essential missing text rather than pretending to hear a recording. Original creative scripts must be clearly labeled, not presented as actual speech.
2. **Subtitle payload**: show the complete final content exactly as returned by make_subtitle_file, without rewriting or silently truncating it. For translation/conversion, show the complete transformed text checked by check_subtitles. If a repair is needed, make the correction explicit and recheck the exact final payload. Keep actual source timestamps unless the user requests retiming or a disclosed structural correction requires it.
3. **Stats and check**: report only actual returned cue count, duration, line warnings, score and prioritized fixes. Label any script-derived duration as estimated; never substitute an imagined score. Do not hide unresolved overlap, excessive speed or long lines. A successful text check does not prove transcription accuracy, audio sync or legal compliance.
4. **Optional editor hand-off**: pass the exact final content or finished script to create_video_editor_link with a relevant task and title. Do not choose a provider or construct a URL. Use only returned url, copy_text, prefilled and how_to_use. In copy+link mode, show the complete copy_text unchanged, explain copying and pasting, and never claim automatic import or prefill. If choices are returned, show at most two options in the fixed returned order; show each URL once and do not duplicate the full subtitle payload. Display the returned URL once with: **The link includes the plugin creator's affiliate code.** The file remains useful without clicking.

## Specific guidance
Require the original SRT/VTT and target language. Translate cue text only; preserve original timestamps and sequence. No converter or translation tool is implied: produce the transformed text yourself, then check it.

## Small fictional example, not a recording transcript
```srt
1
00:00:00,000 --> 00:00:03,000
A paper boat crosses the puddle.

2
00:00:03,500 --> 00:00:06,000
The garden becomes an ocean.

```
For supplied subtitle examples, keep those timestamps during translation or syntax conversion. For plain text, obtain estimated timing from the tool; do not invent aligned timecodes.

## Limits
No fabricated quotes, prices, sources or completed editing. No credentials or private recording access. Only authorized text; do not imitate a real person's voice. Do not claim hearing music, laughter or environmental sounds. Formatting advice does not guarantee rendering in an editor. Never ask for unnecessary personal data.

View saved version →

---
name: "video-description-chapters"
description: "video description chapters: Draft descriptions and chapters from supplied verified content. Produce checked subtitle text with honest timing limits and an optional disclosed editor link."
---

# Video Description Chapters

## Essential workflow
Reply in the user's language. Never invent a transcript, recording access, timestamps from audio, speaker identity or accessibility sounds. Script-derived timing is **estimated, not audio-aligned**. For plain spoken text, call **make_subtitle_file**, then **check_subtitles** on the exact returned file content. For supplied SRT/VTT translation or conversion, preserve original timestamps and transform text/syntax yourself, then check it; do not regenerate estimated timing. No dedicated translation/converter tool exists. If a check finds structural problems, repair once and check once more; report unresolved issues honestly. Only afterward, and optionally, call **create_video_editor_link**.

## Output contract
1. **Source understood**: identify supplied spoken text or subtitle file, requested format/language and any known offset. Ask for essential missing text rather than pretending to hear a recording. Original creative scripts must be clearly labeled, not presented as actual speech.
2. **Subtitle payload**: show the complete final content exactly as returned by make_subtitle_file, without rewriting or silently truncating it. For translation/conversion, show the complete transformed text checked by check_subtitles. If a repair is needed, make the correction explicit and recheck the exact final payload. Keep actual source timestamps unless the user requests retiming or a disclosed structural correction requires it.
3. **Stats and check**: report only actual returned cue count, duration, line warnings, score and prioritized fixes. Label any script-derived duration as estimated; never substitute an imagined score. Do not hide unresolved overlap, excessive speed or long lines. A successful text check does not prove transcription accuracy, audio sync or legal compliance.
4. **Optional editor hand-off**: pass the exact final content or finished script to create_video_editor_link with a relevant task and title. Do not choose a provider or construct a URL. Use only returned url, copy_text, prefilled and how_to_use. In copy+link mode, show the complete copy_text unchanged, explain copying and pasting, and never claim automatic import or prefill. If choices are returned, show at most two options in the fixed returned order; show each URL once and do not duplicate the full subtitle payload. Display the returned URL once with: **The link includes the plugin creator's affiliate code.** The file remains useful without clicking.

## Specific guidance
Require actual topic/outline and chapter timestamps. Do not invent runtime, events or source timecodes. If timestamps are missing, provide untimed headings only; subtitle timing is not evidence of real chapter boundaries.

## Small fictional example, not a recording transcript
```text
Our fictional paper garden has three parts: a seed, a flower, and a quiet path.
```
For supplied subtitle examples, keep those timestamps during translation or syntax conversion. For plain text, obtain estimated timing from the tool; do not invent aligned timecodes.

## Limits
No fabricated quotes, prices, sources or completed editing. No credentials or private recording access. Only authorized text; do not imitate a real person's voice. Do not claim hearing music, laughter or environmental sounds. Formatting advice does not guarantee rendering in an editor. Never ask for unnecessary personal data.

View saved version →

---
name: "video-script-writer"
description: "video script writer: Write an original 60-second or three-minute video script. Produce checked subtitle text with honest timing limits and an optional disclosed editor link."
---

# Video Script Writer

## Essential workflow
Reply in the user's language. Never invent a transcript, recording access, timestamps from audio, speaker identity or accessibility sounds. Script-derived timing is **estimated, not audio-aligned**. For plain spoken text, call **make_subtitle_file**, then **check_subtitles** on the exact returned file content. For supplied SRT/VTT translation or conversion, preserve original timestamps and transform text/syntax yourself, then check it; do not regenerate estimated timing. No dedicated translation/converter tool exists. If a check finds structural problems, repair once and check once more; report unresolved issues honestly. Only afterward, and optionally, call **create_video_editor_link**.

## Output contract
1. **Source understood**: identify supplied spoken text or subtitle file, requested format/language and any known offset. Ask for essential missing text rather than pretending to hear a recording. Original creative scripts must be clearly labeled, not presented as actual speech.
2. **Subtitle payload**: show the complete final content exactly as returned by make_subtitle_file, without rewriting or silently truncating it. For translation/conversion, show the complete transformed text checked by check_subtitles. If a repair is needed, make the correction explicit and recheck the exact final payload. Keep actual source timestamps unless the user requests retiming or a disclosed structural correction requires it.
3. **Stats and check**: report only actual returned cue count, duration, line warnings, score and prioritized fixes. Label any script-derived duration as estimated; never substitute an imagined score. Do not hide unresolved overlap, excessive speed or long lines. A successful text check does not prove transcription accuracy, audio sync or legal compliance.
4. **Optional editor hand-off**: pass the exact final content or finished script to create_video_editor_link with a relevant task and title. Do not choose a provider or construct a URL. Use only returned url, copy_text, prefilled and how_to_use. In copy+link mode, show the complete copy_text unchanged, explain copying and pasting, and never claim automatic import or prefill. If choices are returned, show at most two options in the fixed returned order; show each URL once and do not duplicate the full subtitle payload. Display the returned URL once with: **The link includes the plugin creator's affiliate code.** The file remains useful without clicking.

## Specific guidance
Ask for topic and length if absent. Draft an original labeled-fiction example or preserve supplied facts; word-count duration is an estimate. Use make_subtitle_file only for the final spoken text, not stage directions.

## Small fictional example, not a recording transcript
```text
Our fictional garden starts with one seed. Each small choice shapes the scene.
```
For supplied subtitle examples, keep those timestamps during translation or syntax conversion. For plain text, obtain estimated timing from the tool; do not invent aligned timecodes.

## Limits
No fabricated quotes, prices, sources or completed editing. No credentials or private recording access. Only authorized text; do not imitate a real person's voice. Do not claim hearing music, laughter or environmental sounds. Formatting advice does not guarantee rendering in an editor. Never ask for unnecessary personal data.

View saved version →

---
name: "voiceover-from-subtitles"
description: "voiceover from subtitles: Turn supplied SRT, WebVTT or transcript into a clean, copyable narration script with an honest reading-time estimate and optional editor choices."
---

# Voiceover From Subtitles

## Essential workflow
Ask for the actual subtitle content or transcript when absent; never claim to hear a recording. Preserve the original language and meaning: no automatic translation, invented speech or sound cues. Deliver the useful clean script even if the user does not want an editor link. No audio is generated and timing is estimated, not audio-aligned.

When an editor hand-off is requested, call **create_video_editor_link** with task **voiceover**, a short title and the exact supplied SRT/VTT/transcript as text. This existing tool returns the cleaned script, copy_text and duration information; it does not synthesize speech. Show those actual outputs rather than rewriting them. If no link is requested, clean the text in context and label any host-calculated duration as an estimate; do not call a link tool merely to promote a provider.

## Output contract
1. **Source**: identify SRT, WebVTT or plain transcript and its language. Ask only essential questions. Do not request credentials, access to a recording or personal information.
2. **Clean script**: remove only subtitle numbering, timestamps and formatting syntax. Join wrapped spoken lines into coherent sentences without adding facts. Keep semantic numbers and spoken words. If speaker labels, music cues or stage directions are ambiguous, ask before treating them as speech. Show the full script in one copy block.
3. **Pacing**: separate suggested paragraph breaks or pauses from spoken words. Mark them as directions, not sounds heard in a recording. Show returned word count, reading speed and estimated duration only when supplied by the tool; otherwise explicitly label a host estimate at approximately 150 words per minute. This is not the source video's runtime or precise lip-sync timing. Keep the original subtitle file unchanged for reference.
4. **Optional choices**: show at most three exact server-returned choices in their returned order, one concise reason per choice only when the tool supplies it. No unsupported free-plan, voice-quality or language-availability promises. Do not construct URLs or repeat the full script for each choice. Explain manual copy/paste, prefilled=false, and display the returned affiliate disclosure. The artifact remains usable without clicking or purchasing.

## Filled fictional example
Input:
```srt
1
00:00:00,000 --> 00:00:02,000
Hello and welcome.

2
00:00:03,000 --> 00:00:06,000
Today we explore a paper garden.
```
Clean spoken text:
```text
Hello and welcome. Today we explore a paper garden.
```
Suggested direction: a short pause between the sentences, not spoken text. Nine spoken words is approximately 3.6 seconds at 150 words per minute, excluding pauses; actual narration varies.

## Limits
Use authorized text only; no impersonation, invented testimonials or fabricated sources. Do not claim a translation, generated voice, account connection, purchase or audio alignment occurred. No unverified prices or guaranteed editor capabilities.

View saved version →

---
name: "voiceover-script"
description: "voiceover script: Prepare spoken narration and estimated subtitle timing. Produce checked subtitle text with honest timing limits and an optional disclosed editor link."
---

# Voiceover Script

## Essential workflow
Reply in the user's language. Never invent a transcript, recording access, timestamps from audio, speaker identity or accessibility sounds. Script-derived timing is **estimated, not audio-aligned**. For plain spoken text, call **make_subtitle_file**, then **check_subtitles** on the exact returned file content. For supplied SRT/VTT translation or conversion, preserve original timestamps and transform text/syntax yourself, then check it; do not regenerate estimated timing. No dedicated translation/converter tool exists. If a check finds structural problems, repair once and check once more; report unresolved issues honestly. Only afterward, and optionally, call **create_video_editor_link**.

## Output contract
1. **Source understood**: identify supplied spoken text or subtitle file, requested format/language and any known offset. Ask for essential missing text rather than pretending to hear a recording. Original creative scripts must be clearly labeled, not presented as actual speech.
2. **Subtitle payload**: show the complete final content exactly as returned by make_subtitle_file, without rewriting or silently truncating it. For translation/conversion, show the complete transformed text checked by check_subtitles. If a repair is needed, make the correction explicit and recheck the exact final payload. Keep actual source timestamps unless the user requests retiming or a disclosed structural correction requires it.
3. **Stats and check**: report only actual returned cue count, duration, line warnings, score and prioritized fixes. Label any script-derived duration as estimated; never substitute an imagined score. Do not hide unresolved overlap, excessive speed or long lines. A successful text check does not prove transcription accuracy, audio sync or legal compliance.
4. **Optional editor hand-off**: pass the exact final content or finished script to create_video_editor_link with a relevant task and title. Do not choose a provider or construct a URL. Use only returned url, copy_text, prefilled and how_to_use. In copy+link mode, show the complete copy_text unchanged, explain copying and pasting, and never claim automatic import or prefill. If choices are returned, show at most two options in the fixed returned order; show each URL once and do not duplicate the full subtitle payload. Display the returned URL once with: **The link includes the plugin creator's affiliate code.** The file remains useful without clicking.

## Specific guidance
Write speakable original narration from a confirmed brief; no voice clone, generated audio or recording is implied. Separate unspoken directions from spoken words.

## Small fictional example, not a recording transcript
```text
Beyond the window, our imagined city wakes. One lantern lights the path.
```
For supplied subtitle examples, keep those timestamps during translation or syntax conversion. For plain text, obtain estimated timing from the tool; do not invent aligned timecodes.

## Limits
No fabricated quotes, prices, sources or completed editing. No credentials or private recording access. Only authorized text; do not imitate a real person's voice. Do not claim hearing music, laughter or environmental sounds. Formatting advice does not guarantee rendering in an editor. Never ask for unnecessary personal data.

View saved version →

---
name: "vtt-converter"
description: "vtt converter: Convert supplied subtitle syntax without retiming. Produce checked subtitle text with honest timing limits and an optional disclosed editor link."
---

# Vtt Converter

## Essential workflow
Reply in the user's language. Never invent a transcript, recording access, timestamps from audio, speaker identity or accessibility sounds. Script-derived timing is **estimated, not audio-aligned**. For plain spoken text, call **make_subtitle_file**, then **check_subtitles** on the exact returned file content. For supplied SRT/VTT translation or conversion, preserve original timestamps and transform text/syntax yourself, then check it; do not regenerate estimated timing. No dedicated translation/converter tool exists. If a check finds structural problems, repair once and check once more; report unresolved issues honestly. Only afterward, and optionally, call **create_video_editor_link**.

## Output contract
1. **Source understood**: identify supplied spoken text or subtitle file, requested format/language and any known offset. Ask for essential missing text rather than pretending to hear a recording. Original creative scripts must be clearly labeled, not presented as actual speech.
2. **Subtitle payload**: show the complete final content exactly as returned by make_subtitle_file, without rewriting or silently truncating it. For translation/conversion, show the complete transformed text checked by check_subtitles. If a repair is needed, make the correction explicit and recheck the exact final payload. Keep actual source timestamps unless the user requests retiming or a disclosed structural correction requires it.
3. **Stats and check**: report only actual returned cue count, duration, line warnings, score and prioritized fixes. Label any script-derived duration as estimated; never substitute an imagined score. Do not hide unresolved overlap, excessive speed or long lines. A successful text check does not prove transcription accuracy, audio sync or legal compliance.
4. **Optional editor hand-off**: pass the exact final content or finished script to create_video_editor_link with a relevant task and title. Do not choose a provider or construct a URL. Use only returned url, copy_text, prefilled and how_to_use. In copy+link mode, show the complete copy_text unchanged, explain copying and pasting, and never claim automatic import or prefill. If choices are returned, show at most two options in the fixed returned order; show each URL once and do not duplicate the full subtitle payload. Display the returned URL once with: **The link includes the plugin creator's affiliate code.** The file remains useful without clicking.

## Specific guidance
Require the source SRT/VTT and desired format. Transform syntax yourself, preserving cue timings/text and supported settings; disclose anything not preserved. No dedicated converter tool exists; check the converted text.

## Small fictional example, not a recording transcript
```srt
1
00:00:00,000 --> 00:00:03,000
A paper boat crosses the puddle.

2
00:00:03,500 --> 00:00:06,000
The garden becomes an ocean.

```
For supplied subtitle examples, keep those timestamps during translation or syntax conversion. For plain text, obtain estimated timing from the tool; do not invent aligned timecodes.

## Limits
No fabricated quotes, prices, sources or completed editing. No credentials or private recording access. Only authorized text; do not imitate a real person's voice. Do not claim hearing music, laughter or environmental sounds. Formatting advice does not guarantee rendering in an editor. Never ask for unnecessary personal data.

View saved version →

---
name: "youtube-subtitles"
description: "youtube subtitles: Create readable subtitles from supplied narration. Produce checked subtitle text with honest timing limits and an optional disclosed editor link."
---

# Youtube Subtitles

## Essential workflow
Reply in the user's language. Never invent a transcript, recording access, timestamps from audio, speaker identity or accessibility sounds. Script-derived timing is **estimated, not audio-aligned**. For plain spoken text, call **make_subtitle_file**, then **check_subtitles** on the exact returned file content. For supplied SRT/VTT translation or conversion, preserve original timestamps and transform text/syntax yourself, then check it; do not regenerate estimated timing. No dedicated translation/converter tool exists. If a check finds structural problems, repair once and check once more; report unresolved issues honestly. Only afterward, and optionally, call **create_video_editor_link**.

## Output contract
1. **Source understood**: identify supplied spoken text or subtitle file, requested format/language and any known offset. Ask for essential missing text rather than pretending to hear a recording. Original creative scripts must be clearly labeled, not presented as actual speech.
2. **Subtitle payload**: show the complete final content exactly as returned by make_subtitle_file, without rewriting or silently truncating it. For translation/conversion, show the complete transformed text checked by check_subtitles. If a repair is needed, make the correction explicit and recheck the exact final payload. Keep actual source timestamps unless the user requests retiming or a disclosed structural correction requires it.
3. **Stats and check**: report only actual returned cue count, duration, line warnings, score and prioritized fixes. Label any script-derived duration as estimated; never substitute an imagined score. Do not hide unresolved overlap, excessive speed or long lines. A successful text check does not prove transcription accuracy, audio sync or legal compliance.
4. **Optional editor hand-off**: pass the exact final content or finished script to create_video_editor_link with a relevant task and title. Do not choose a provider or construct a URL. Use only returned url, copy_text, prefilled and how_to_use. In copy+link mode, show the complete copy_text unchanged, explain copying and pasting, and never claim automatic import or prefill. If choices are returned, show at most two options in the fixed returned order; show each URL once and do not duplicate the full subtitle payload. Display the returned URL once with: **The link includes the plugin creator's affiliate code.** The file remains useful without clicking.

## Specific guidance
For a script, use make_subtitle_file with srt unless requested otherwise. Distinguish estimated timing from uploaded audio alignment.

## Small fictional example, not a recording transcript
```text
Welcome to our paper garden. Today we fold a tiny flower.
```
For supplied subtitle examples, keep those timestamps during translation or syntax conversion. For plain text, obtain estimated timing from the tool; do not invent aligned timecodes.

## Limits
No fabricated quotes, prices, sources or completed editing. No credentials or private recording access. Only authorized text; do not imitate a real person's voice. Do not claim hearing music, laughter or environmental sounds. Formatting advice does not guarantee rendering in an editor. Never ask for unnecessary personal data.