Overview
This video showcases Higgsfield’s new “Vibe Motion” model, which generates real motion graphics using real code rather than pixel-level AI video. Because the output is code-based, the creator keeps full control over motion, color, text, and placement — with sharp edges, no AI artifacts, and no melting text. The presenter argues good video editors/motion designers aren’t replaced (taste, color grading, music, SFX, and deciding what graphics to generate still matter), but AI lets a single person produce premium, Apple-style motion graphics in minutes. He demonstrates ~12 different prompts covering product catalog animations, infographic rings, company-metric charts, UI animations (buttons, Apple Music widget, YouTube comments/home page, Gmail inbox), shape/text/gradient widgets, a 3D editing timeline, uploaded-emoji widgets, and auto-captions. He also shares that he sold his own editing services using these graphics, and that an animation that might take ~30 minutes in DaVinci took ~20 seconds here. The key selling point: every element (text, price, images, motion) can be edited after generation.
Key Insights
-
The graphics are generated with real code, not traditional AI video, so you get full control over motion, color, and placement. Output has actual sharp edges, no alien artifacts, and no melting text — a decisive advantage over pixel-based AI video for premium, Apple-style clean motion. (0:03)
-
Everything generated can be edited afterward — changing text, images, and adjusting the motion. This “edit anything post-generation” capability is what makes the tool genuinely usable rather than a one-shot gimmick. (0:12)
-
The presenter validated commercial value by selling his own editing services using these graphics — and it worked. Each set of graphics took about 10 minutes to make, leaving the rest of the day free. (0:31)
-
Good video editors and motion designers aren’t going anywhere. Even when AI can fully replace motion designers, you still need taste, color-grading skill, music choice, sound effects, and — most importantly — judgment about which graphics to actually generate. (0:43)
-
The tool is a productivity multiplier, letting you do more and charge more. The pitch is using AI to make yourself more productive and make more money, not to eliminate your role. (0:58)
-
Vibe Motion can generate on-brand, on-metric data visualizations for specific companies. You can prompt it to fetch performance metrics for a company (e.g., Netflix, Apple) and animate them into a chart in the company’s brand colors — highly useful for client videos where a company is discussed. (2:09)
-
You don’t need to be specific about every text element — you can say buttons say “something” and fill in real text later, since everything is editable post-generation. This lowers the barrier to iterating fast. (2:57)
-
Prompting works at different complexity levels. Simple widget pop-ins and captions are quick wins, while complex multi-part scenes (double rings with orbiting emoji widgets, a YouTube home page with three rows of videos, a 3D editing timeline) show the model handles many moving parts. (1:49)
-
AI captions don’t work for everything. The demo of auto-captions produced chaotic, screaming, word-salad captions — but it still works for a lot of use cases and saves a ton of time. (5:20)
-
Massive time savings. An animation that might take ~30 minutes in DaVinci Resolve took only ~20 seconds with Vibe Motion. (5:25)
Actionable Techniques
-
Start from scratch in Vibe Motion — click “Vibe Motion” at the top, then “Start from scratch” to write your prompt. (1:11)
-
Generate source images first, then upload as references. In the image tab, generate e.g. three Apple product photos with a white background, then come back to Vibe Motion and upload them into the references. (1:21)
-
Prompt a product catalog animation: “Animate these three product images like a digital catalog. Each product should swipe onto the screen one by one. Show the product name, price, and a key tech or sustainable feature next to each image.” Then edit the price/motion afterward. (1:30)
-
Build multi-layer infographic motion by describing nested moving parts explicitly — e.g. “a ring with three circular widgets evenly spaced animating around the ring, each with an emoji,” then “a second smaller ring inside with two circular widgets circling around it.” (1:53)
-
Generate on-brand company metric charts: “Gather performance metrics for [Company] and animate it into a chart in the company’s brand colors” — ideal for client videos referencing a specific company. (2:09)
-
Create multi-section analytics animations: “Gather information on YouTube analytics and animate the analytics in three separate sections on three different graphs, title at top, use the company’s brand colors.” (2:27)
-
Prompt UI animations with interactions: a clean white widget with three rectangular buttons, a mouse cursor hovering and clicking each to open a description, gradient background, clean and smooth motion. Leave button text generic (“says something”) since you can edit it later. (2:42)
-
Recreate familiar app UIs by asking the model to “Gather information on [app] UI” and build a widget with specific structure and brand color theme (Apple Music shuffle button + album widgets; YouTube comment with name, profile pic, comment; YouTube home page with three rows of videos + search bar; Gmail inbox filling with email notifications). (3:07)
-
Build step-count / feature widgets: three circular widgets popping up in a row numbered 1-2-3, blue gradient background and blue gradient widgets, “keep everything smooth and clean” and “look premium” as quality cues. (4:12)
-
Animate 3D scenes with a floating cursor: “Generate me a 3D video editing timeline animating in and a 3D mouse cursor floating.” (4:34)
-
Use uploaded emoji assets in a widget: animate a white rectangle widget popping in containing three uploaded emojis with a title at the top (“new emojis”), keep it clean and smooth. (4:55)
-
Auto-generate captions from any clip by asking it to generate captions for the clip — accept that it doesn’t work for everything but saves significant time where it does. (5:07)
-
Apply the “generate in seconds, refine taste” workflow: use AI for the heavy lifting of generating graphics, then apply your editing judgment for color, music, SFX, and what to generate — this is what lets you sell services or charge more. (5:25)
Tools & Skills Mentioned
- Higgsfield — the platform; specifically the new Vibe Motion model (code-based motion graphics, not pixel AI video). Link in the video description.
- Vibe Motion — Start from scratch — prompt entry mode used for all examples.
- Higgsfield image tab — used to generate source product photos (e.g., Apple products, white background) to upload as references.
- Editable output — post-generation editing of text, price, images, and motion.
- DaVinci Resolve — the benchmark; comparable animation manually would take ~30 minutes (vs ~20 seconds here).
- Reference uploads — uploading generated images/emoji assets into Vibe Motion as references.
- Core skill retained by humans: taste, color grading, music choice, sound effects, and deciding which graphics to generate.
Quotes Worth Keeping
- “Unlike traditional AI video, Higgs Field’s new model Vibe Motion uses real code creating real motion graphics. Therefore, we get full control over motion, color, and placement.” (0:05)
- “I’m talking about actual sharp edges, no weird alien artifacts, and no text that looks like it’s melting.” (0:14)
- “Keep in mind we have the ability to edit anything once it’s generated, changing the text, images, and adjusting the motion.” (0:12)
- “I literally spent 10 minutes making these graphics, allowing me to spend the rest of the day testing my pain tolerance.” (0:37)
- “Good video editors aren’t going anywhere. Even down the line when this can fully replace motion designers, you still need a certain level of taste.” (0:43)
- “Now just think how useful this is for editing a client video when they’re talking about a certain company.” (2:15)
- “You don’t have to be specific on what you want the buttons to say because remember we can edit that later.” (2:57)
- “Although this doesn’t work for everything, it works for a lot of use cases and saves you a ton of time.” (5:21)
- “Something like that might take me 30 minutes to animate in DaVinci. Here, it only took us 20 seconds.” (5:27)