# Mad Monkey — video generation rules

Updated 2026-09-20 · v2 · For text-to-video tools (Veo, Sora, Runway, Kling and
successors). These rules override any conflicting instruction in a brief.

## Format first

- **9:16 vertical (1080×1920) is the default.** TikTok, Reels, Stories.
- 16:9 (1920×1080) only for slides, screens and YouTube.
- Keep key action in the centre band — top ~200px and bottom ~320px are
  platform-UI zones on vertical.

## The logo rule (same as images — non-negotiable)

Never let a video model generate, animate or "reveal" the Mad Monkey logo, any
monkey mascot or any wordmark. No exceptions. Logos and end cards are added in
the edit, using official files only.

## End frames — use the official ones

Every vertical video signs off with THE official end frame (1080×1920, hold
1–2s, hard cut — no fades):

https://design.madmonkeyhostels.com/end-frames/mm-endframe-black.png

Do not rebuild, re-type or re-animate it.

## What video models are for

Footage plates only: b-roll energy, motion backgrounds, transitions. Bake in
**no text, no logos, no captions** — type is added in the edit as real
Montserrat. Prefer REAL footage of real guests wherever it exists; generated
people must never be presented as real guests or staff.

## Look

Same as photography: faces, sweat, water, sunset, motion, confetti. Saturation
pushed, mid-tones warm, blacks crushed, handheld energy. No drone-only empty
landscapes, no corporate smoothness, no slow-motion luxury-brand pacing.
Cut fast. Loud beats polished.

## Prompt skeleton

"Vertical 9:16 handheld footage, [subject: backpackers mid-action], golden hour
/ neon night, saturated colour, crushed blacks, fast motion, no text, no logos,
no watermarks, documentary energy" — length 6–20s per clip; build sequences
from multiple short clips rather than one long generation.
