The Automation Boost Video Pack

Build an award-winning cinematic "30 second explainer" website for [YOUR_WORKFLOW_NAME].

VISUALS — use the seedance_2.5 model on Higgsfield MCP. Settings: mode omni_reference, resolution 720p, 1600, generate_audio false. First generate ONE hero image with GPT Image 2 — [DESCRIBE MATERIAL, COLOUR, FINISH, LIGHTING] on a clean reflective surface — and pass that job id as start_image.

Then generate ONE single clip at duration 30 (this is the whole point of 2.5 — the entire journey in one unbroken take, no chaining, no seams, no continuity drift).

1. OPEN (duration 3) – keyframe: ultra-wide shot of the whole workflow: nodes, connectors, the data flowing left to right. Camera pulls back. Motivation: "here is what we are building." Continuous smooth camera motion throughout, no jitter.

2. NODE_ONE (duration 8) – keyframe: zoom in on first node, magnify the inputs, the transformation logic ("this is where you map your data"). Camera: slow dolly forward, one continuous shot, no cuts, no shakiness.

3. NODE_TWO (duration 8) – keyframe: the second node magnifies into the frame, inputs flow in from node one, outputs flow right to node three. Camera: smooth lateral pan, perfect seamless turntable. The process iterates; no camera shake, photorealistic travel cinematography, soft push motion throughout, no jitter.

4. THE_OUTPUT (duration 8) – keyframe: final node outputs the clean transformed data, display counter spinning up in the top corner, the data stream flowing into a clean, bright database, the final result shown clearly, one continuous shot, ending on the payload perfectly still at the end, one continuous shot.

WEBSITE — slice the clip to 180 frames at 1600px and scroll-scrub it as one canvas sequence across four pinned chapters, so scrolling fills the visitor from the top left into the process flow data, one column and one row wide. Then normal sections: suites grid below, a LIVE_STATUS_COUNTER (one real live endpoint hooked to your actual workflow), and a RESERVE_A_DEMO section with date and party-size selectors. Design: clean MONTSERRAT_CAPS, one dark theme, file grain. Copy tone: confident, direct. Launch on localhost and verify every scroll animation works before telling me it's done.

Build an award-winning cinematic "screen-record tutorial" website for [YOUR_COURSE_NAME].

VISUALS — seedance_2.5 on Higgsfield MCP. Settings: mode omni_reference, resolution 720p, 1600, 1612, generate_audio false. First generate ONE hero screenshot of your n8n canvas at a nice zoom level ([DESCRIBE NODES, LAYOUT, COLOURS]) using GPT Image 2, then pass that as start_image on the tutorial clip.

Then generate ONE single clip at duration 28:

1. HERO_REVEAL (duration 4) – keyframe: the n8n canvas enters from the left side of frame, zooms to fill the screen, nodes glowing gently. One continuous shot, camera motion: ultra-gentle forward push, no shaking.

2. THE_FIRST_NODE (duration 7) – keyframe: the first node magnifies center-screen, the config panel slides open, the inputs highlight in sequence as you describe them on screen. Camera: slow dolly forward, one continuous shot, no fast cuts, no jitter. Motivation: "here is your trigger."

3. THE_MIDDLE_LOGIC (duration 8) – keyframe: zoom out slightly, two nodes in frame, the data wire between them glows and pulses, the transformation is applied, the output flows right and down. Camera: slow lateral pan across the nodes, continuous smooth motion, one take. Motion: "see how the data transforms at each step."

4. THE_FINAL_OUTPUT (duration 9) – keyframe: the workflow ends at the output node, the send-to-Slack step executes (if applicable), the message appears in a Slack window overlay in the bottom right, the final data is displayed, one continuous shot, ending very still, camera motion whatsoever. Motivation: "and this is your result landing in the right place."

WEBSITE — slice to 180 frames at 1600px and scroll-scrub the tutorial as one canvas sequence. Then build a real tutorial page with two columns: left column is the scroll-scrub video, right column is step-by-step text (with code blocks for n8n JSON if applicable). A COPY_CONFIG button exports the workflow JSON. A DEPLOY_NOW button triggers your n8n API to duplicate the workflow into the visitor's account. Design: dark theme, ROBOTO_MONO for code, BEBAS for headings. Copy tone: conversational, encouraging. Launch on localhost and verify the workflow export works before telling me it's done.

Build an award-winning cinematic "AI avatar explainer" website for [YOUR_PERSONAL_BRAND].

VISUALS — seedance_2.5 on Higgsfield MCP. Settings: mode omni_reference, resolution 720p, 1600, 1612, generate_audio false. First generate ONE portrait of [YOUR_AVATAR_DESCRIPTION] in a black studio ([DESCRIBE LIGHTING, MATERIALS, POSE, MOOD]) using GPT Image 2, then pass that as start_image on every clip. This is what keeps your avatar's face identical throughout all shots.

Then generate ONE single clip at duration 26:

1. ENTRANCE (duration 5) – keyframe: the avatar sits at rest, composed, lit by clean studio light. Camera slowly orbits around the avatar's shoulders in a black-walled studio, very slow dolly forward, one continuous shot, no cuts, camera motion smooth and gentle, no shakes. Motivation: "here is your guide."

2. THE_PITCH (duration 9) – keyframe: the avatar gestures as it speaks (if you are comfortable with motion), or stays perfectly still if you prefer. Behind the avatar, abstract automation diagrams flow softly in the background (nodes, wires, data pulses). Camera: locked or very slow push in, continuous still shot, no jitter. Motivation: "listen to what I am building."

3. THE_DETAIL (duration 7) – keyframe: the avatar nods slightly, the background diagrams become more complex, a glowing accent bar pulses in the bottom third (use your brand colour here). Camera: locked, continuous shot, motion smooth. Motivation: "this is the mechanism."

4. THE_CLOSE (duration 5) – keyframe: the avatar stands, the background fades to black, the avatar is lit only by studio light, one final confident expression, ending perfectly still, one continuous shot, no motion at the end.

WEBSITE — slice the clip to 180 frames at 1600px and scroll-scrub it across the full hero. Then build a real site: the scroll-scrub video occupies the full above-the-fold, then below it a TEXT_SECTION with a three-column grid of your three strongest value props (one line each). Then a FAQ accordion, then a BOOK_A_DEMO section with an elegant inline calendar. Design: dark AUTOMATION_BOOST branding, your brand colour accent throughout, serif fonts for body (elegance), BEBAS for headings. Copy tone: confident, direct, no fluff. Launch on localhost and verify the scroll and calendar work before telling me it's done.

Build an award-winning cinematic "location tour" website for [YOUR_RESTAURANT_OR_SERVICE].

VISUALS — seedance_2.5 on Higgsfield MCP. Settings: mode omni_reference, resolution 720p, 1600, 1612, generate_audio false. First generate ONE wide-angle interior shot of [DESCRIBE SPACE, MOOD, LIGHTING, MATERIALS] using GPT Image 2 — then pass that as start_image on every clip. This keeps the space consistent throughout the sequence.

Then generate ONE single clip at duration 28:

1. THE_ENTRANCE (duration 6) – keyframe: the camera enters through the front door, pulls back slightly, reveals the full dining room. Lighting: warm, inviting, candles on tables, soft overhead light. Camera: slow continuous push inward, no cuts, photorealistic travel cinematography, one smooth motion throughout.

2. THE_DETAIL (duration 7) – keyframe: zoom in on a table setting, the plating, a cocktail being set down, the menu open to a signature dish ([DESCRIBE INGREDIENTS, PRESENTATION]). Camera: slow lateral pan across the table, continuous smooth motion, one take, no jitter.

3. THE_KITCHEN (duration 8) – keyframe: transition to the kitchen or prep area (if you choose), hands working, ingredients in motion, the craftsmanship visible. Camera: slow overhead or side pan, continuous motion, photorealistic, no fast cuts.

4. THE_ATMOSPHERE (duration 7) – keyframe: pull back to a table with guests (or a wider room shot), ambient light, warm tones, the final moment: one guest raising a glass, the room bathed in that moment. Camera: slow push forward, ending very still, one continuous shot, no shake.

WEBSITE — slice to 180 frames at 1600px and scroll-scrub the hero cinematic as the full-width above-the-fold video. Below: MENU_SECTION (if applicable), a RESERVE_TABLE panel with live date / time / party-size selectors, PHOTOS_GRID (with your actual photos as gallery), HOURS_AND_LOCATION (embedded map), a REVIEWS section pulling from Google. Design: dark elegant theme, brand colour accents (golds, warm tones if your aesthetic is warm; cool greys if it is modern), serif fonts for body, BEBAS for headings. Copy tone: welcoming, descriptive, sensory language. Launch on localhost and verify the reservation form works before telling me it's done.

Build an award-winning cinematic "product showcase" website for [YOUR_PRODUCT].

VISUALS — seedance_2.5 on Higgsfield MCP. Settings: mode omni_reference, resolution 720p, 1600, 1612, generate_audio false. First generate ONE hero shot of [YOUR_PRODUCT] with [DESCRIBE MATERIAL, COLOUR, FINISH, LIGHTING] on a black reflective floor using GPT Image 2 — then pass that as start_image on every clip. This is what keeps your product identical throughout every shot.

Then generate ONE single clip at duration 30:

1. HERO_REVEAL (duration 8) – keyframe: the product rotates slowly while the camera orbits it, studio lights sweeping across the surface, the material texture revealed. Camera: smooth 360-degree rotation, one continuous shot, no cuts. Motivation: "here is the thing you are about to buy."

2. MATERIAL_CLOSE (duration 8) – keyframe: extreme close-on the product's key detail — stitching, engraving, the gem, the mechanism — lit to show off the craftmanship. Camera: slow dolly forward, one continuous shot, no shakiness. Motivation: "see the quality up close."

3. IN_HAND (duration 8) – keyframe: the product sits in a hand (or in use), the scale revealed, the action shown, the detail glowing under studio light. Camera: static or very slow push, continuous shot, one take, photorealistic travel cinematography, soft lighting.

4. THE_REVEAL (duration 6) – keyframe: pull back, the product on its own against black, one continuous shot, no cuts, one final moment of stillness.

WEBSITE — slice the clip to 180 frames at 1600px and scroll-scrub it as the full-width hero. Then build: SPECS_TABLE (dimensions, materials, weight), PRICING_SECTION (variants, stock status), ADD_TO_CART button, REVIEWS_SECTION pulling from Shopify or your store, SHIPPING_AND_RETURNS, FAQ. Design: dark premium theme, minimal clutter, your brand colour as accent, serif for body, BEBAS for headings. Copy tone: sophisticated, technical where relevant, conversational where it sells. Launch on localhost and verify the add-to-cart flow works before telling me it's done.