<aside> 🛠

Choose the Exact Cut and Framing Inside Each Candidate Take — The Shorts Planner

The decision-making brain of the intelligent-shorts chain. Video EDL Planner takes the candidate takes (from Video Analysis) plus constraint measurements (from Video Constraint Map) and figures out exactly where to cut inside each take, what framing to use per moment, and how to pace the finished short. Set a target duration, pick a pace (Original / A little faster / Much faster), optionally bias toward "interesting" time ranges — and let the planner emit the edit plan that Video EDL then converts into the segments the rest of the shorts chain renders.

</aside>

What it does

Video EDL Planner is the shorts planner — it consumes the candidate highlight takes from Video Analysis and the frame-level constraint measurements from Video Constraint Map, then makes editorial decisions: which parts of each take to use, exactly where to make each cut (within a configurable search granularity), how to frame each moment (respecting the constraint map's subject positions and safe crops), and how many cuts to make (driven by the chosen pace). The output is a coherent edit plan — the intermediate representation that Video EDL turns into the segments JSON + per-clip metadata the rendering chain consumes.

Problem it solves

• Turns takes + constraints into a plan – Video Analysis says "these are the candidate moments"; Video Constraint Map says "here's where the subject is and what crops are safe"; Video EDL Planner decides the actual edit
• Duration-aware editing – Aims to hit a target duration (default 45 s), naturally shortens when the takes can't fill the budget
• Pace control – One dropdown (Original / A little faster / Much faster) changes both the number of cuts and their length — matches the editorial energy to the platform / content
• Content-aware bias – Point the planner at "interesting" time ranges of the source (e.g., "0:55~3:20") and it will pull more material from that region
• Cut-search granularity – Sub-second search resolution finds better cut points than fixed-boundary trimming; balances quality against arithmetic cost
• Composition-safe cuts – Respects the constraint map's subject positions so no cut ever loses the subject mid-frame
• Deterministic outputs – Same inputs + same settings = same plan, so the pipeline is reproducible and auditable
• Separates decisions from rendering – Planner decides what to do; Video EDL formalizes it; render tools handle how to do it — each stage is testable and swappable

Input/Output

<aside>

• Input: Candidate takes + constraint measurements
  • Takes (top input, yellow-gold port): The candidate highlight takes to plan against — typically produced upstream by Video Analysis (or a similar candidate-source tool)
  • Constraints (second input, orange port): The per-frame constraint measurements from Video Constraint Map — subject positions, candidate crop windows, caption-safe flags </aside>

<aside>

• Output: Edit plan (orange port)
  • Format: Machine-readable plan describing which slice of each take to use, exactly where to cut, what framing to apply, and how the finished short is composed
  • Consumed by: Video EDL (converts the plan into the segments JSON + per-clip metadata the rendering chain expects) </aside>

Configuration Options

• How long the finished short may run: Target duration cap for the planned short.
  • Default: 45 seconds
  • UI: Slider (main node view)
  • Behavior: The planner aims to come in under this length; when the candidate takes can't supply enough footage, the plan falls short rather than padding
  • Reference points:
    • 15 – 30 s – Short-form TikTok / Reels / Shorts pacing
    • 45 s (default) – Standard short-form range; enough time to develop a moment but still snackable
    • 60 – 90 s – Longer social / YouTube Shorts maximum
    • 120+ s – Multi-highlight compilations, still short-form but with more content

• How fast the short cuts: Pace / cutting rhythm. Note: Pace changes both the number of cuts the planner makes and their individual lengths — the two are coupled to keep the pacing coherent.

  Option	Behavior	Best For
  Original	Preserves the natural rhythm of each take; fewer, longer cuts	Interview / podcast content, contemplative pacing, storytelling where the moment needs to breathe
  A little faster (default)	Moderate uptempo pacing; more cuts, slightly shorter each	Standard short-form social; energetic without being frenetic
  Much faster	Aggressive cutting; many shorter cuts	Meme / comedy / high-energy content; TikTok / Reels where kinetic pacing = retention