Screenshot Punch P01
0.0s
Previews are illustrative demonstrations of the technique — not your footage. Your real edit is produced later by Claude Code.
P01 · Proof · beginner · 1.5–8s
A real screenshot drops in over the speaker at a precise scale, lands with a soft overshoot and a deep shadow, holds long enough to read, then clears.
The moment the speaker makes a claim, the supporting screenshot punches into frame: it enters slightly oversized, settles to its final width with a short overshoot and casts a layered shadow that separates it from the footage. The background dims or blurs just enough to make the capture the only thing to read. Because it lands on the exact word that references it, the screenshot reads as evidence rather than decoration.
Try the parameters
Local path Claude Code can read. Never uploaded.
Final width as % of frame width.
Best used for
- — Showing a result or analytics screenshot
- — Backing up a claim mid-sentence
- — Tutorial checkpoints
- — Reaction and commentary videos
Assets
- ● Source video required
- ● Screenshots required
- ○ Transcript with timing optional
- ○ Sound effects optional
Pairs well with
If assets are missing
If no screenshot is supplied, do not recreate, mock up or 'illustrate' the evidence — omit this beat and list it as a missing asset, or ask the user for the capture. If the screenshot is too low-resolution to read, show it smaller with a crop around the relevant area (keeping enough context to stay honest) and flag it.
Accessibility
Hold ≥ 1.5s plus 1/3s per word of relevant text in the capture. Dim the background enough for the screenshot to stand out; describe what the screenshot shows in the voiceover or captions so it is not visual-only information.
Individual prompt
updates with the parameters above# CES style instruction — P01 Screenshot Punch You are an experienced motion designer and video editor working in Claude Code. Apply the **P01 Screenshot Punch** technique to the supplied footage as specified below. **B. Project configuration:** 1080×1920, 9:16, 30 fps · Source video: not provided — ask the user for the path before starting · Video language (on-screen text & captions): English **Proof** — The moment the speaker makes a claim, the supporting screenshot punches into frame: it enters slightly oversized, settles to its final width with a short overshoot and casts a layered shadow that separates it from the footage. The background dims or blurs just enough to make the capture the only thing to read. Because it lands on the exact word that references it, the screenshot reads as evidence rather than decoration. ### P01 Screenshot Punch - **When:** ≈ 00:00.00 → 00:03.00 (≈ derived — adjust to the nearest natural beat in the footage) - **Layer / track:** V2+ (overlay above the picture) - **What it does:** A real screenshot drops in over the speaker at a precise scale, lands with a soft overshoot and a deep shadow, holds long enough to read, then clears. - **Parameters:** Screenshot file: (not set); Duration: 3 s; Entrance: Drop + punch (from above, overshoot); Screenshot width: 82%; Landing tilt: -2 deg; Shadow depth: 60%; Background treatment: Dim footage 45%; Easing: Back out (overshoot) - **Steps:** 1. Place screenshot from the project assets (the user's supplied screenshot; if none is present, stop and ask for it — do not substitute a mock-up) on an overlay layer above the source footage from 00:00.00 to 00:03.00 (1080×1920 @ 30fps). Use only the supplied material, exactly as captured: do not retype, re-word, re-colour, crop out context that changes meaning, or 'clean up' numbers, names or dates. If anything must be hidden (private handles, emails, order numbers, faces of bystanders), blur or block it visibly so viewers can tell it was redacted. 2. Scale it to 82% of frame width (886px), preserving its aspect ratio; if it would exceed 70% of frame height, fit by height instead. Never upscale beyond 200% of its native pixels — if it would, show it smaller rather than blurry. Round corners by ≈ 2.5% of its width. 3. Entrance over 14 frames: enter from 12% of frame height above its rest position at scale 1.12 and land at scale 1.0, using a back-out curve with slight overshoot (cubic-bezier(0.34, 1.56, 0.64, 1)), while rotating from -5.0° to a resting tilt of -2°. Opacity reaches 1 within the first 3 frames. 4. Shadow: two stacked drop shadows — a tight contact shadow (y offset 0.6% of frame width, blur 1.2%, 31% black) and a soft ambient one (y offset 3%, blur 8%, 37% black). Add a 1px 10% white inner stroke so dark screenshots separate from dark footage. 5. Dim the underlying footage to 55% brightness over 9 frames as the screenshot enters, restoring it on exit. 6. Land the punch on the exact word in the transcript where the speaker references the screenshot. Hold it fully still for at least 2.3s — long enough to read every word that matters (≈ 3 words per second); extend the range rather than cutting it short. 7. Exit over 8 frames (scale 1.0 → 0.96, opacity → 0, ease-in). Verify on a phone-sized render that the key text in the screenshot is legible at 82% width. - **Assets:** Source video (⚠ MISSING), Screenshots (⚠ MISSING) · optional: Transcript with timing, Sound effects - **Fallback:** If no screenshot is supplied, do not recreate, mock up or 'illustrate' the evidence — omit this beat and list it as a missing asset, or ask the user for the capture. If the screenshot is too low-resolution to read, show it smaller with a crop around the relevant area (keeping enough context to stay honest) and flag it. **G. Asset integrity and restrictions:** - Do not create or present invented testimonials, comments, quotes, reviews or endorsements. - Do not fabricate numbers, metrics, charts or results. Data visualisations use only the supplied data. - Do not create fake screenshots or UI and present them as real. Illustrative graphics must be clearly illustrative. - Do not attempt unsupported 3D reconstructions or perfect subject separation without a mask; use the documented 2D fallbacks instead. **I. Quality verification:** - Check output dimensions and aspect ratio with ffprobe — they must match the configuration exactly. - Check the output frame rate and that the duration is within ±0.5 s of the target. - Check audio/video synchronisation at the start, middle and end (lip-sync and SFX hits within 1 frame). - Inspect frames at every beat boundary for text clipping, overflow and safe-zone violations. - Confirm the output file exists and plays — never claim a render is complete without an actual output file. - Report any missing asset or unimplemented effect explicitly.
More proof styles
Use proof whenever you make a claim. Every proof style requires authentic, supplied assets — never fabricated ones.
- P02Comment SpotlightOne real comment is lifted out of the feed onto a clean card, the rest of the frame dims, and the key phrase is marked with a highlighter sweep.
- P03Metrics Count-UpA verified number ticks up from a meaningful starting value to its exact final figure in large tabular numerals, with its source label underneath.
- P04Before and After WipeTwo aligned real visuals share the frame while a clean divider line sweeps across, revealing the after state over the before, then rests on a split.
- P05Search-to-ResultA search bar types the exact query character by character, a short loading beat follows, and it resolves into the real captured result card.
- P06Receipt StackSeveral real proof captures deal onto the frame one by one, each fully readable before the next lands, building a tidy stack of receipts.
- P07Annotation MarkerA hand-drawn circle, arrow, underline or box draws itself onto the exact part of a real screenshot, with the rest of the image gently dimmed.
- P08Timeline EvidenceA chronological track draws across the frame and real dated events pop onto it one at a time, each with its date label and supporting capture.
- P09Quote Card RevealA supplied quotation is set in elegant type on a card, revealed line by line behind large quote marks, with the author and source fading in last.
- P10App UI BreakdownThe camera zooms from a real full-screen capture into one interface element, dims the rest, and a callout line labels what that element actually does.