Introduction
The Simple Edits Creators—including Single Cut, Double Cut, and Triple Cut—are designed to streamline video editing by extracting and delivering only the verbatim, sentence-complete segments from transcription needed for impactful, concise video outtakes. These creators do not script additional framing content or editorial text, but rather, focus entirely on chunk selection, thus enabling instant video assembly based precisely on the spoken material.
How It Works
Input Section
Each Creator accepts:
Transcription: An srt-encoded transcript with chunk indices and spoken content.
User Input: Optional instructions that can guide segment focus, terminology, or tone.
Theme Sensitivity and Focus
- If User Input exists, it is parsed to extract key terms, themes, or preferences, which are used to guide chunk selection.
- If no User Input is present, the model identifies salient topics and keywords from the frequency, emphasis, or explicit statements in the transcription.
Chunk Selection Protocol
In all variants:
- Sentence Boundary Enforcement: Each selection begins at a sentence start (previous chunk ends with ., !, or ?, current chunk starts with uppercase/quote) and ends with a sentence terminator.
- Chunks must always be consecutive; never skip or fragment sentences.
- For each cut, include whichever consecutive chunk range is necessary to maintain grammatical and narrative completeness—even if this slightly exceeds word count.
- Typical target is 30–45 words per TH selection, up to 70 if that is required for sentence closure. Always prioritize sentence integrity.
- Selections emphasize:
- Clear definitions, claims, or thematic introductions.
- Key reasoning, consequences, actions, or recommended steps.
- Standalone insights, recommendations, and natural spoken moments.
Block Logic
Single Cut: Delivers one array (TH1) representing the most thematically relevant and complete sentence(s).
Double Cut: Delivers two arrays (TH1, TH2), each representing a strong, distinct sentence or concept that together remain cohesive and thematically unified.
Triple Cut: Delivers three arrays (TH1, TH2, TH3) depicting a logical progression (set-up, insight, action/implication) or three linked but standalone points.
User Input Role
- User Input can direct focus, word choice, tone, and the weighting of segments, ensuring results match editorial or client needs.
Writing Section
There is no script writing—output is exclusively:
- Chunk Indices: For each section (one to three), the model outputs an array of chunk indices.
- Call to Action: At the end, a short CTA is included, sourced from system settings or defaulting to "Thank you for watching." in the appropriate language.
Best Use Scenarios
These Creators are ideal for:
Rapid Social Edits: Quickly clipping out key statements or moments from longer interviews or presentations.
Sizzle Reels or Event Recaps: Stringing together the sharpest takeaways or most memorable moments.
Internal Comms/Training: Highlighting essential messages for team or stakeholder sharing.
User Guide
Step-by-Step Instructions
- Provide the Transcription: Upload or paste the srt transcript (chunked input).
- Set User Input: Optionally specify themes, keywords, or tone.
- Run the Creator: The model parses, applies theme focus, and returns the relevant chunk indices and CTA.
- Review and Deploy: Use the index output for precise video clip exports or automation.
Tinker with it
Direct the Angle: Choose user-provided keywords or focus for each video, refining the nature of each cut.
Adjust for Channel: Change the preferred number of cuts or CTA for platform, campaign, or format requirements.