Brand Film Prompt Structure
A highly detailed prompt template for generating a 55-second brand film using Opus 5.5. It includes specific inputs (logo, archival clips, song), direction (visual style, banned elements), and structure (cold open, past, bridge, button, end).
Prompt
<inputs>
Ask me for: a brand name, its logo (SVG or high-res PNG), one line of what it becomes in 5 years (not today's features), one coined word the film should teach (e.g. "signal"), a public-domain archival clip of the old way of doing it (https://t.co/t5EaRxnryq Prelinger), an archived web page from ~2000 where people asked for things (Wayback Machine), a royalty-free song ~100-120 BPM with a real drop (Mixkit, free for commercial use), and a free ElevenLabs key for the narrator (premade voices only).
</inputs>
<direction>
An a16z x Content Rewards brand film, ~55s. It sells one feeling and one idea, never features. Real archival and real web pages, never stock footage, never product screens, never fake UI. One hero object: a 3D button with the logo embossed into it, the ONLY saturated colour in the film. Everything else white, silver, cream and archival black. Small white captions that light up word by word on the narrator's timestamps. Film texture: fresh grain every frame, halation on bright edges, soft vignette, warm lifted blacks. The brand appears at the end, never before.
Banned: stock footage, dashboards, chat UIs, feature tours, cards without real profile pictures, wobbly "worm" graph lines, random glitches, any hold longer than 1.5s, purple, rainbow gradients, orange on anything but the button.
</direction>
<structure>
Cold open (0-8s): the archival clip plays on a glowing old CRT TV with its real dialogue, captions word by word.
The past (8-17s): push into the real 2000 web page, a cursor lands on one real "wanted" post, old pages strobe one per frame, a counter ticks EVERY frame up to today's scale. Narrator: "People have always asked the internet for what they need." / "Today, there are five hundred million posts a day."
Bridge (17-30s): a galaxy of real old asks. One ask snaps to a white outline with a chip of the coined word ("We call that a signal."). Archival of people doing it by hand on the TV. Cut on each word: "Door by door. Call by call. Scroll by scroll." The narrator never stops talking between acts.
Button (30s): the button rises out of black, rim light first, colourless. "[Brand] finds the signals." A flat grey line: "this is how you find customers by hand". Hard cut to a button macro, the colour comes alive, the music drop lands on this exact frame. "and this... is with [Brand]": a steep lit curve, counters changing every frame, a glowing sine with a comet head the camera chases, then platform strikes (X, Reddit, LinkedIn, YouTube, Instagram, TikTok) each with its own counter, then the galaxy again with glowing platform logos: "Every signal on the internet." Hyperspace.
End: the tagline word by word, then "[Brand] is live.", ONE cream frame on the kick, then logo + URL in silence.
</structure>
<build>
1. One HTML file, 1920x1080. Everything is computed from time inside seek(t): no CSS transitions, no timers, no Math.random (seeded hash), no state between frames.
2. Button: three.js MeshPhysicalMaterial (clearcoat 1, roughness ~0.4), ACES filmic, cool + warm rim spots, room env map. Emboss the real logo by displacing a dense disc with a blurred mask of the logo, folded into the rim so corners don't show. Brushed-metal base.
3. Counters: linear on the film's global clock, so they keep running through cutaways. Tabular numerals, fixed unit label.
4. Captions: <=7 words, one idea per shot, hand-placed line breaks, one text layer on screen at a time. Grey to white over 0.2s on the word timestamps (ElevenLabs /with-timestamps).
5. Glitch = 1-2 frames, only on a real cut, always with a click. Readable things stay on screen >= 8 clean frames.
6. Render with Playwright per frame (Chromium with --use-angle=swiftshader so WebGL renders headless), finish in numpy: fresh grain per frame, halation, vignette ~0.65x at the corners, 1px chroma fringe with EDGE-CLAMPED shifts. Pipe PNGs into ffmpeg, h264, 24fps.
7. Mix: narrator 10-15dB over the music, duck -18dB while speaking, hits on the exact action frame, -14 LUFS.
</build>
<gotchas>
If the narrator pauses, the visuals must not stop. If a counter freezes, the render is broken. If the button isn't the most memorable thing in the film, start over.
</gotchas>
How to use this prompt
- 1
Copy the full prompt above.
- 2
Open a platform that supports Claude Opus 5.5, such as YouMind, and paste the prompt in.
- 3
Swap the subject, style, or details to fit your idea, then generate.
This is a free AI prompt from YouMind's prompt library. Explore thousands more video prompts, all free to copy and adapt.
Explore more video promptsMore from Claude Opus 5.5
High-End Product Video Prompt
@zero
Business Explainer Video Prompt
@Alex Prompter
Kinetic Product Promo Motion Design Prompt
@Martijn Verbove
Product Video Generation Prompt
@Benza → LicitaCerta.com.br
Professional SaaS Product Launch Video
@Dinesh S
Website Promo Video with Remotion and Python Music
@卡尔的AI沃茨
All Video Categories
01
Use Cases
More Prompt Features
AI Library
AI prompt search
Let AI search through tens of thousands of prompts. Filter by model, time range, keywords, and sort by engagement — views, bookmarks, reposts and more.
Vision Tools
Image to Prompt
Turn any photo into a detailed AI image prompt. The free image to prompt converter analyzes composition, style, and lighting so you can recreate any look in seconds.
Other Prompt Features
Browse by Model
Browse by Media
Built for creators. Free forever.
YouMind is the AI creative copilot trusted by millions of creators worldwide. Every prompt here is curated to help you create better, faster.