0:00 / 0:00

Comment “AI” for the link 🤯🚀 What a great time to be alive. Whether you’re generating thumbnails from scratch or taking inspiration from others, @lovart.ai is amazing for both ⚡ And the best thing is It’s super easy to use just head over to Lovart, upload a few inspiration thumbnails, and write in plain English how you want to take inspiration from them and create yours, and edit them if needed. #ai #genai #aithumbnails #lovart #lovartai

Simone Ferretti

@sferro21

INSTAGRAM · 2026-03-01Source

534likes

665comments

Remix This

Recreate with Kling 3

Make your own AI viral video

Prompt

GLOBAL LOCK: A fast-paced vertical 9:16 creator-tool Reel, approximately 39 seconds, built from alternating talking-head footage and bright white SaaS interface screens. Keep the host consistent across the whole timeline: white male in his late 20s to early 30s, slim build, side-parted brown hair, clean-shaven, wearing a dark charcoal or black crew-neck top, speaking into a large black front-address microphone, framed in a rounded-corner lower panel with warm amber-brown studio background and soft flattering key light. The second visual world is a clean Lovart interface shown mostly on white backgrounds with black text, light-gray cards, and thumbnail canvases. The third visual world is a set of YouTube-style thumbnails focused on male self-improvement / money / outreach / rules content, often with bold yellow or blue text, a male presenter portrait on the right, chart graphics, arrows, and high-contrast click-oriented design. Maintain aggressive jump-cut pacing every 1 to 2 seconds, crisp UI readability, no clutter outside the interface elements, and one energetic male speaker throughout with close-mic dry audio, clear English cadence, excited but practical tone, lips visible during host shots.

[00:00-00:03] Open on a white Lovart-branded background showing three example thumbnails in a row: a bold yellow money thumbnail, a red-and-white net worth thumbnail, and a numbered bottle-style comparison graphic. The host appears in a rounded talking-head box below, raising one finger as he starts the hook. Audio: single male host announces that this is a better or easier way to create thumbnails, or that the tool changes how thumbnail ideation works.

[00:03-00:07] Cut between the host and more sample thumbnails, including a finance/investing guide style card with a male face, dollar annotations, and editing instructions, then an outreach chart thumbnail labeled 4 TYPES OF OUTREACH. The host keeps pointing upward, matching the beat of each new thumbnail. Lighting on host stays warm; thumbnails stay bright, white-dominant, and highly legible.

[00:07-00:11] Move into the Lovart file/document view on a dark gray workspace with multiple imported thumbnail references arranged in a grid. Show the host below continuing to talk while the interface demonstrates that several inspiration images can be loaded at once. Audio: same speaker explains that you can upload inspiration thumbnails and work from them directly.

[00:11-00:16] Transition to the white Lovart prompt workspace labeled Design is easier with Lovart. Show a large prompt field and a card-based layout where attached references appear on the right or top. Cursor focus should imply a natural-language command is being entered. Audio: host explains that you simply tell the system in plain English how to transform or take inspiration from the reference thumbnail.

[00:16-00:22] Show a sequence of prompt cards with explicit edit instructions, such as recreating the second attached thumbnail, placing the host face from the first image, or keeping the main text while changing the face. The host continues in the lower frame, gesturing rhythmically. The interface should feel beginner-friendly and not overly technical. Audio: one speaker emphasizes simplicity and speed, with phrases about using plain English and uploading inspiration thumbnails.

[00:22-00:27] Reveal a zoomed view of the edited prompt result and the generated thumbnail outputs. Use variants around a self-improvement thumbnail reading 5 Rules 1. No phone after 10pm, including slight typography or layout changes across versions. Show how the face, title placement, and color block stay recognizable while the design gets remixed. Audio: host underscores that the tool can iterate thumbnail ideas faster than building each one manually.

[00:27-00:33] Expand the result set into multiple derivative thumbnails, such as 7 Mantras 1. No food after 10pm or revised 5 Rules variants, then display UI controls that suggest choosing among versions or continuing to refine. Keep the white interface clean and uncluttered, with the host below maintaining eye contact. Audio: host frames the value in terms of scaling from one idea to multiple ideas and faster testing.

[00:33-00:36] Show the final polished YouTube-style thumbnail on a dark background with a clear title treatment, then cut to a comparison-style CTA row at the bottom featuring labels such as 3x faster, 10x faster, STOP, Manual, and AI. The host intensifies his delivery to set up the conversion. Audio: same speaker makes the benefit explicit: more thumbnail ideas in less time.

[00:36-00:39] End on a bold CTA screen reading COMMENT "AI" while the host remains in the lower panel finishing his line. Use strong white text with yellow AI accent on a dark card, and retain the polished thumbnail examples nearby so the CTA remains tied to the promised outcome. Audio: punchy final ask to comment AI for the link, no fade.

NEGATIVE PROMPT: avoid unreadable thumbnail text, avoid distorted charts or arrows, avoid broken facial features on sample thumbnail portraits, avoid muddy interface screenshots, avoid random extra UI panels, avoid low-resolution white backgrounds, avoid incorrect Lovart branding, avoid host identity drift, microphone deformation, warped hands during finger-pointing, robotic speech, lip-sync mismatch, harsh compression, clipping, or CTA text other than COMMENT "AI".

How sferro21 Made This Lovart Thumbnail Generator AI Video - and How to Recreate It

This Reel is a clean creator-growth pitch for thumbnail ideation at scale. Simone Ferretti frames Lovart as a practical way to generate and remix YouTube thumbnails by uploading inspiration images and then giving plain-English edit instructions. The visuals are intentionally simple: white Lovart pages, imported thumbnail references, lightweight document cards, prompt boxes, and a sequence of output thumbnails that look like they belong to money, productivity, outreach, and self-improvement channels. The host stays in a warm talking-head frame with a microphone, which keeps the product demo from feeling sterile. The outcome is easy to understand even if you miss some spoken words: start with one thumbnail idea, attach examples, describe what you want changed, and get multiple usable variants faster than manual design. That is a strong creator promise because thumbnails sit at the top of the growth funnel. If you can multiply thumbnail options quickly, you can test faster, publish faster, and improve click-through rates without rebuilding each concept from scratch.

What You're Seeing

The opening uses familiar thumbnail language

The first frames show click-oriented YouTube thumbnails with bold text, money cues, face cutouts, and simple chart graphics. That immediately tells the viewer what category of problem the Reel is solving.

The host functions like a fast translator

He is not deeply designing on screen. He is translating the product into creator language: upload references, describe changes in plain English, get more ideas. That keeps the message accessible for smaller creators.

The product proof is workflow-based, not hype-based

You see imported thumbnails, the prompt workspace, instruction cards, and the resulting thumbnail variants. That is enough to make the flow credible without slowing the Reel down.

The examples stay in one niche on purpose

Most output thumbnails revolve around money, rules, habits, outreach, and self-improvement. That consistency makes the demo easier to follow because the viewer is not trying to parse too many different content genres.

The CTA lands after enough evidence has accumulated

By the time Comment "AI" appears, the viewer has already seen source thumbnails, instruction prompts, and multiple output variations. The conversion ask comes after proof, not before it.

Shot-by-Shot Breakdown

Time range	Visual content	Shot language	Lighting & color tone	Viewer intent
0:00-0:03 (estimated)	Three example thumbnails on a clean white Lovart background	Static gallery-style opener with host panel below	Bright white UI, bold saturated thumbnail colors	Make the use case instantly legible
0:03-0:07 (estimated)	More thumbnail examples including finance and outreach concepts	Quick swaps between sample thumbnails and host	White background, heavy yellow and blue title blocks	Signal thumbnail variety and creator relevance
0:07-0:11 (estimated)	Lovart document/file view with imported thumbnail references	Dark workspace insert with grid layout	Dark gray interface with small preview cards	Prove reference images can be uploaded and organized
0:11-0:16 (estimated)	Design is easier with Lovart prompt workspace	Interface zoom into prompt and attachment layout	Clean white canvas with soft gray cards	Show low-friction natural-language workflow
0:16-0:22 (estimated)	Detailed prompt cards explaining how to reuse attached thumbnails and faces	Card-by-card instruction reveal	White UI with black text and subtle controls	Demonstrate that the tool understands edit intent
0:22-0:27 (estimated)	Generated output thumbnails around the “5 Rules” concept	Fast comparison sequence of similar variants	Blue-green gradients, big white text, male presenter portrait	Show iteration speed and output usefulness
0:27-0:33 (estimated)	More derivative thumbnails like “7 Mantras” and alternate “5 Rules” versions	Variation stack with minor layout and wording changes	Consistent niche palette and click-driven composition	Sell quantity and testing potential
0:33-0:39 (estimated)	Performance/efficiency comparison and COMMENT “AI” CTA	End-card style conclusion with host finishing the pitch	Darker end-card accents over white product context	Turn product proof into comment conversion

Why It Went Viral

It targets a growth bottleneck every creator understands

Thumbnail creation sits close to views, CTR, and publishing speed. That makes the topic broadly relevant to YouTube creators, social educators, AI tool enthusiasts, and even short-form creators who understand the value of packaging.

The workflow is easy to explain and easy to imagine using

This is not an abstract promise about “better creativity.” It is a visible, specific workflow: upload a few thumbnails, describe what you want in plain English, generate more versions. Simplicity like that performs well because it reduces skepticism.

The demo sells leverage instead of design theory

The Reel never turns into a long design lesson. It keeps the framing practical: more ideas, faster variations, easier iteration. For growth-minded creators, that is usually more compelling than a deep lecture about typography or composition.

The thumbnail niche helps the post feel commercially useful

The example outputs are not random art. They are click-oriented, monetizable thumbnail concepts. That gives the Reel immediate business value and makes saves more likely.

The CTA matches the level of viewer curiosity

Once viewers see the workflow and output range, it is natural for them to want the link. Asking them to comment AI is a low-friction next step that keeps engagement on-platform.

Platform Signals

The first second is visually obvious

You instantly know the video is about thumbnails because the screen is filled with obvious YouTube-style examples. That reduces explanation cost and improves stop rate.

The edit pattern alternates proof types

The Reel cycles through example thumbnails, product workspace, natural-language prompts, generated results, and CTA. That mix helps preserve retention because each new cut answers a different viewer question.

It has strong save and share logic

Creators can save it as a tool reference or send it to a teammate with a simple message: “This is the thumbnail workflow I was talking about.” That is a good distribution pattern for creator-tool content.

5 Testable Viral Hypotheses

Hypothesis 1: Thumbnail examples increased first-second comprehension

Observed evidence: the Reel opens on recognizable thumbnail cards. Mechanism: viewers understand the topic before they fully process the narration. Replication: lead with the finished assets, not the dashboard.

Hypothesis 2: Plain-English prompting reduced intimidation

Observed evidence: the interface shows natural-language instructions rather than technical parameter settings. Mechanism: viewers believe they could use the tool themselves. Replication: highlight human-readable prompts in your demo.

Hypothesis 3: Reusing one niche made the results easier to compare

Observed evidence: many outputs stay within the self-improvement / money thumbnail genre. Mechanism: viewers can spot the iterations quickly because the variables are controlled. Replication: keep one content niche throughout your demo.

Hypothesis 4: Quantity framing increased perceived value

Observed evidence: the later frames suggest 3 ideas, 5 ideas, 10 ideas, and faster production. Mechanism: creators associate more options with more testing power and better growth outcomes. Replication: explicitly frame the output in terms of option volume.

Hypothesis 5: The comment CTA converted because the reward was obvious

Observed evidence: the caption and ending both point viewers toward commenting AI for the link. Mechanism: a one-word comment is easy, immediate, and connected to a specific benefit. Replication: keep the CTA short and clearly reward-linked.

How to Recreate It

Step 1: Pick a packaging problem with direct growth value

Thumbnail ideation works well because creators immediately understand why more options matter.

Step 2: Gather 3 to 5 reference thumbnails

Choose examples from the same niche so the transformations feel coherent and easy to judge.

Step 3: Keep your host setup simple

One warm talking-head setup with a microphone is enough to humanize the tutorial and keep pacing under control.

Step 4: Show the upload step clearly

Viewers need to believe they can start from inspiration images they already have, not from a blank creative void.

Step 5: Write prompts the viewer could actually copy

Use plain English instructions like “recreate the second attached thumbnail with my face from the first attached image.” This is exactly the kind of phrasing that lowers adoption resistance.

Step 6: Show multiple output variants, not one hero image

The point of the tool is iteration speed, so one output is not enough evidence.

Step 7: Tie the result to growth language

Frame the benefit around more ideas, faster testing, and easier publishing rather than just prettier design.

Step 8: End with a CTA linked to access

If you want comments, promise something concrete: a link, a workflow, or the tool name plus instructions.

Step 9: Publish with platform-native framing

Keep the pacing fast, the proof visual, and the CTA low-friction so the Reel behaves like social content, not like a help-center video.

Growth Playbook

3 opening hook lines

1. This is the easiest way I have found to make better thumbnails faster.

2. If you already have one thumbnail idea, this tool can turn it into ten.

3. Stop making every thumbnail from scratch if you can do this instead.

4 caption templates

Template 1: Thumbnail ideation gets a lot easier when you can upload references and describe changes in plain English. This is why Lovart is so useful for creators. Comment AI for the link.

Template 2: If you are still building every thumbnail concept manually, you are probably moving slower than you need to. This workflow turns one idea into multiple testing options fast. Comment AI.

Template 3: The best part of this tool is not just the design quality, it is the option volume. More variants means faster testing and faster publishing. Comment AI for the link.

Template 4: Upload a few inspiration thumbnails, explain what you want changed, and you can generate fresh thumbnail ideas way faster. If you want to try it, comment AI.

Hashtag strategy

Broad: #AI #GenAI #YouTubeTips #ContentCreation. These widen the discovery net.

Mid-tier: #AIThumbnails #ThumbnailDesign #CreatorTools #YouTubeGrowth. These align more closely with the actual use case in the Reel.

Niche long-tail: #Lovart #LovartAI #ThumbnailIdeation #YouTubeThumbnailWorkflow. These target viewers already evaluating thumbnail-generation tools.

FAQ

Why does this Reel feel useful immediately?

Because the opening frames already show the exact kind of assets creators want more of: clickable thumbnails.

What is the most important proof shot in this video?

The natural-language prompt cards, because they make the workflow feel accessible instead of technical.

Why do the examples stay in one niche?

Keeping the subject matter consistent makes the iterations easier to compare and the value easier to grasp.

How can I copy this without making it look generic?

Use references from one niche, show your actual instructions, and stack multiple variants so the speed advantage becomes obvious.

What makes the CTA work here?

The viewer has already seen the full workflow and only needs a simple path to get the link.

Is this better for YouTube creators than general AI art creators?

Yes, because the output shown in this Reel ties directly to packaging, testing, and click-through performance.