Created by STORYBOARD
Step 1: Prepare Main Footage
Prepare the main video before entering Higgsfield.
- Video length: under 10 seconds.
- Ratio: 9:16 or 16:9 only.
- Footage should be clean, readable, and ready for motion graphics overlay.
Step 2: Design Text Motion Reference
Generate a still text-motion reference image first. This image defines the typography style, color direction, layout behavior, and graphic attitude for the video prompt.
- Use GPT Image 2 to create the text motion reference.
- Keep the design clear: sports typography, speed streaks, fitness data, and strong readable hierarchy.
Prompt (copy/paste):
{
"task": "Create one single vertical 9:16 reference motion graphic image only. One image only.",
"format": {
"ratio": "9:16 vertical",
"background": "pure solid black",
"color_palette": {
"allowed": ["pure white", "vivid sports blue"],
"forbidden": ["gold", "pink", "green", "red", "any extra colors"]
}
},
"style": [
"bold condensed italic sports typography",
"premium athletic editorial campaign",
"clean kinetic typography",
"sharp horizontal speed streaks",
"fast motion smear trails",
"small simple blue line icons",
"clean stat layout",
"modern sports commercial look"
],
"main_typography": {
"words": "RUN / WITH / CONTROL",
"hierarchy": {
"RUN": "large at the top, bold white italic type",
"WITH": "in the middle, large vivid blue outline word",
"CONTROL": "large below, bold white italic type"
},
"effect": "fast horizontal white and blue speed streaks trailing from the right side of the letters"
},
"secondary_text": {
"text": "KEEP THE PACE",
"placement": "underneath the hero text",
"style": "smaller spaced-out vivid blue typography with thin blue horizontal line accents"
},
"stats_section": {
"layout": "clean lower stat layout with small blue line icons and white numbers",
"exact_stats": [
"DISTANCE 5.35 KM",
"GOAL 10 KM",
"AVG PACE 5:28 /KM",
"AVG HEART RATE 156 BPM",
"TIME 29:17",
"CALORIES 233"
],
"support_line": "BREATHE • FOCUS • HOLD (small, at the bottom)"
},
"layout": [
"keep everything centered and vertical",
"clean poster-like hierarchy",
"lots of black negative space",
"composition feels like a clean reference board for motion typography",
"large stacked hero text on top, two rows of compact stats below"
],
"negative": [
"no dashboard UI",
"no sci-fi HUD",
"no scan visuals",
"no grunge texture",
"no background patterns",
"no full scene or person",
"no extra stats",
"no misspelled text",
"no grid",
"no multiple versions in one image"
]
}Step 3: Open Higgsfield
Open higgsfield.ai and start a new video generation workflow.
Step 4: Select Gemini Omni Flash
Select the Gemini Omni Flash model before uploading the files, so the workflow is already in the correct model mode.
Step 5: Upload References
- Upload MainVideo
- Upload TextMotionRef
Step 6: Enter Text Motion Prompt
The prompt should explain how the text motion is built, not just what text appears on screen.
Prompt structure. The text motion prompt should mainly reference these parts:
- Original footage rule: keep the existing video unchanged.
- Visual style: follow the uploaded typography reference image.
- Video timing: split the motion by time range, such as 0-3s, 3-6s, and 6-9.25s.
- Text groups: define which words or data appear in each phase.
- In / hold / out rhythm: every text group should animate in, hold briefly, then animate out completely.
- Placement: use negative space and avoid blocking the main subject.
- Motion language: speed streaks, masked reveals, wipe-ins, small icon ticks, and clean lock-ins.
Prompt (copy/paste):
{
"task": "Add premium kinetic typography overlay to the existing 9.25s vertical treadmill running video",
"original_footage_rule": "Keep the original footage unchanged: same runner, treadmill, room, lighting, outfit, and camera movement",
"visual_style": {
"reference": "use the uploaded blue-white sports typography reference as the visual style",
"look": [
"bold condensed italic sports type",
"white and vivid blue only",
"clean speed streaks",
"small simple line icons",
"premium athletic editorial look"
]
},
"negative": [
"no full-screen poster",
"no dashboard interface",
"no sci-fi interface",
"no scan visuals",
"no extra colors",
"no texture or grunge",
"no broken text or misspellings",
"do not add FLOW, CADENCE, SPM, or extra stats"
],
"audio": {
"sfx": "clean motion SFX only: soft whooshes, swishes, reveal ticks, lock-in hits, and streak wipe sounds synced to the typography",
"forbidden": ["no music", "no voice", "no ambience"]
},
"core_rule": "The video has 3 visual phases. Use exactly 3 typography groups. Only one group appears at a time. Each group must animate in, hold briefly, then animate out completely before the next group starts. No previous group may remain, overlap, ghost, duplicate, or return by accident.",
"text_data": [
"RUN WITH CONTROL", "KEEP THE PACE",
"DISTANCE 5.35 KM", "GOAL 10 KM", "AVG PACE 5:28 /KM",
"AVG RATE 156", "TIME 29:17", "ENERGY 233",
"BREATHE", "FOCUS", "HOLD"
],
"groups": [
{
"group": 1,
"time": "0.0-3.0s",
"shot": "Wide Opening Shot",
"layout": "large hero typography in the wide negative space",
"sequence": [
"RUN enters first, large white, fast left-to-right speed smear",
"WITH enters second, vivid blue outline, masked wipe reveal",
"CONTROL enters third, large white, strong horizontal wipe and lock-in",
"KEEP THE PACE appears smaller underneath with blue tracking expansion"
],
"exit": "at 3.0s, all Group 1 text exits completely with a clean speed-streak wipe"
},
{
"group": 2,
"time": "3.0-6.0s",
"shot": "Full-Body Mid Shot",
"layout": "compact stat typography, no giant hero text, lower-third or side negative space",
"show_only": ["DISTANCE 5.35 KM", "GOAL 10 KM", "AVG PACE 5:28 /KM"],
"style": "small blue abstract line icons, animate stats in one by one with masked reveals and small ticks",
"exit": "at 6.0s, all Group 2 stats and icons exit completely"
},
{
"group": 3,
"time": "6.0-9.25s",
"shot": "Closer Angle Change",
"layout": "most compact layout, no giant center text",
"show_only": ["AVG RATE 156", "TIME 29:17", "ENERGY 233"],
"style": "stats reveal one by one, hold briefly, then wipe out",
"kinetic_beats": [
"BREATHE enters then exits",
"FOCUS enters then exits",
"HOLD enters, locks in, then exits"
],
"final_lockup": "RUN WITH CONTROL, 5.35 KM • 5:28 /KM • 156 (small, clean, edge-aligned, readable)",
"ending": "smooth blue-white streak wipe or soft fade"
}
],
"motion_style": [
"fast in, clean hold, fast out",
"masked text reveals",
"horizontal speed streaks",
"motion blur trails",
"small icon ticks",
"clean lock-ins"
],
"camera_aware_placement": "large text only in the wide opening, compact text only in the closer ending"
}Full workflow summary
- Prepare a clean main video under 10 seconds, in 9:16 or 16:9.
- Generate a still typography reference image with GPT Image 2.
- Open higgsfield.ai and start a new video workflow.
- Select Gemini Omni Flash before uploading files.
- Upload the main video and the typography reference.
- Write the text motion prompt: timed phases, one text group at a time, in / hold / out, camera-aware placement.
The key is control: the reference image locks the look, the prompt locks the timing.