DEV Community: Lee Stuart

I Accidentally Generated 47 Ad Variants in One Lunch Break (And Only Hated 3 of Them)

Lee Stuart — Wed, 24 Jun 2026 02:08:15 +0000

Let me start at the end: I now have a folder called ad_variants_DO_NOT_DELETE sitting on my desktop with 47 files in it. My client approved 11 of them. I didn't plan any of this.

The Part Where I Explain How I Got Here

It was a Tuesday. I was at my usual spot — that slightly-too-loud independent coffee shop near the canal where the WiFi password is written on a chalkboard in chalk that's been there since 2023. I had 45 minutes between calls. I was supposed to be eating a sandwich.

I wasn't eating the sandwich.

Instead, I was poking around an AI ad creative generator I'd been putting off testing for two weeks. The brief was simple enough: a DTC skincare brand needed creatives for a retargeting campaign. Three platforms, two audience segments, one very opinionated founder who had "a vision."

I'd been doing this kind of work manually for about three years. You know the drill — brief, moodboard, copywriter ping-pong, three rounds of revisions, someone's cousin has feedback. The usual.

The Accidental Experiment

Here's where it gets embarrassing (in a productive way, I promise).

I was trying to generate a single static ad for the 9:16 format. I misread the interface. Instead of selecting "generate 1 variant," I somehow had it set to batch mode — I think I'd been clicking around too fast — and hit confirm while reaching for my coffee.

The tool started generating. And kept generating. And then kept generating some more.

By the time I looked back at the screen, it had produced 12 variants of the same ad — different headline placements, different CTA button positions, different background color temperatures, same core visual. All from one prompt. All in about 90 seconds.

I sat there for a moment. Then I ate some of the sandwich.

What I Actually Learned About AI Avatar Video Ads (The Hard Way)

Here's the thing about AI avatar video ads that nobody really talks about in the breathless "10x your output!" posts: the value isn't in the single output. It's in the variance.

When I was generating manually — or even using AI tools one-at-a-time — I was unconsciously anchoring to my first instinct. First layout. First color choice. First CTA phrasing. I'd iterate, sure, but always from that anchor point. My "variations" were basically the same ad wearing slightly different shoes.

What the accidental batch generation showed me was that the tool had no such anchor. It was genuinely exploring the space. Some outputs were terrible (one had the CTA button in the lower-left corner, which, why). Some were fine. But three of them were genuinely surprising — layouts I wouldn't have arrived at through my normal process.

The founder, who had "a vision," picked two of those three as her top choices.

(I did not tell her they were accidents.)

The A/B Testing Angle Nobody Warned Me About

So I leaned into it. For the next two weeks, I started using the AI ad creative generator specifically as an A/B testing variant machine rather than a "make me an ad" machine. Subtle but important reframe.

The workflow looked roughly like this:

Write one solid creative brief with locked brand elements (colors, logo safe zones, tone)
Generate 8–12 variants per format in batch
Do a fast cull — remove anything with obvious layout failures or off-brand color drift
Present 3–4 survivors per format to the client
Let the actual ad platform data do the rest

The result? We went from "one creative per format, revised three times" to "four creatives per format, tested in market, optimized based on real CTR data." Same budget. Same timeline. More signal.

The part I didn't expect: the client conversation changed. Instead of defending creative choices ("why is the button blue?"), I was presenting options and framing it as a testing hypothesis. Clients love feeling like scientists. File that one away.

Where It Still Falls Apart

I want to be honest here because the "AI replaced my entire workflow" narrative is exhausting and also not true.

Brand color consistency remains genuinely annoying. Even with locked hex codes in the prompt, the tool occasionally drifts — a slightly warmer white here, a slightly more saturated accent there. Not catastrophic, but enough that I do a manual color-check pass on everything before it goes to the client. That's maybe 10 minutes per batch, which is fine, but it's not zero.

Copy-visual tone matching is the other one. The tool is good at layout logic. It's less good at understanding that a brand with a dry, minimalist voice probably shouldn't have a CTA that reads "GLOW UP NOW ✨" even if the visual composition is technically sound. I've started writing CTA copy separately and injecting it rather than letting the generator suggest it.

And the avatar video side of things — when I started experimenting with animated avatar variants rather than static ads — the consistency issues compound. Lip sync on the avatar was fine for short clips (under 15 seconds). Anything longer and there's a subtle drift that I can't quite describe except to say it looks like someone who's had one too many coffees and is trying very hard to appear normal. (I say this with affection. I also look like this sometimes.)

The Folder on My Desktop

Back to those 47 files.

After the skincare campaign, I started doing this for every project — generating more than I needed, culling fast, testing what survived. The DO_NOT_DELETE folder is actually an archive of every variant I've generated across four client projects. I keep it because occasionally I'll be working on something new and think "wait, didn't I accidentally solve this layout problem three months ago?" and I'll go digging.

It's become a weird kind of creative library. One that I built mostly by clicking the wrong button.

I've since tried a few tools in this space. The one I've stuck with for avatar video ad variants specifically is Nextify.ai — mostly because the batch generation interface is explicit rather than something you stumble into by accident, which is, you know, an improvement on my original workflow.

The Part I'm Still Thinking About

Here's what I haven't fully resolved: if the best creative decisions I made in Q1 came from outputs I didn't deliberately design — if the variants that performed best were the ones the tool generated while I was reaching for a sandwich — what exactly was my contribution?

I wrote the brief. I did the cull. I framed the client conversation. I caught the color drift. I fixed the CTA copy.

That's still something. I think.

But honestly — if the tool is generating the variance and the market is selecting the winners, am I a creative director or just a very expensive filter?

14 Observations From Rebuilding My Ad Workflow Around AI Tools (The Unglamorous Version)

Lee Stuart — Wed, 10 Jun 2026 02:11:59 +0000

A friend who does similar work — content production, ad creative, the whole messy stack — sent me a voice note last week. Three minutes long. The gist: "I don't understand why you're still doing voiceovers manually. You're spending four hours on something that takes four minutes now."

She wasn't wrong. She was annoying about it, but she wasn't wrong.

So I spent the better part of two weeks pulling AI Ad Builder tools and AI Text to Speech engines into my existing workflow. Not replacing it. Pulling them in, which is a different and much more friction-filled process than the demos make it look.

Here's what I actually noticed. No rankings. No scores. Just observations, in the order I wrote them down.

1. The workflow friction is front-loaded, not ongoing.

The first three days were genuinely unpleasant. Figuring out where the AI output slots into my existing pipeline — what it replaces, what it feeds into, what it breaks — took longer than I expected. After that, it got easier. The pain is real but it's not permanent. Worth knowing before you quit on day two.

2. AI Text to Speech has a "valley of almost-good" problem.

The voices aren't bad. That's the issue. They're good enough that you keep thinking one more tweak will get them to broadcast quality. It won't. There's a ceiling, and it's lower than the demos suggest. For internal content or low-stakes social ads, fine. For anything where voice is part of the brand identity, you'll feel the gap.

3. Dynamic caption generation is the sleeper feature nobody talks about enough.

I expected to care most about the visual generation. I ended up caring most about dynamic subtitle generation. Accurate, auto-timed captions that adapt to pacing changes — this alone saved me somewhere between 45 minutes and two hours per project depending on length. It's not exciting to write about. It's extremely exciting to experience at 6pm when you're tired.

4. The AI Ad Builder treats "brand voice" as a style parameter. It isn't.

Every tool I tested had some version of a brand voice input. Tone selectors, adjective fields, example copy uploads. The output was consistently adjacent to brand voice, not inside it. Brand voice isn't a style setting. It's accumulated context — years of decisions about what a company doesn't say as much as what it does. AI doesn't have that context. You have to hold it yourself.

5. TTS pacing is harder to control than pitch or tone.

Most AI Text to Speech interfaces give you decent control over speed and emotion. What they don't give you is good control over micro-pacing — the half-beat pause before a key word, the slight acceleration through a list, the breath before a CTA. These are the things that make a voiceover feel like a person made it. SSML tags help but they're fiddly and the results are inconsistent across engines.

6. Generating caption styles that match video tone is still a manual job.

The tool auto-generates the text. It does not auto-generate the right visual treatment for that text. Font weight, color contrast, animation timing, whether the captions sit inside or outside the safe zone — all of that still requires human decisions. The generation handles the what. The how it looks is on you.

7. The best use case I found wasn't the one I went in looking for.

I went in trying to speed up final production. The actual win was in early-stage iteration. Using the AI Ad Builder to rough out six or seven structural variants of an ad before I committed to any direction — that changed how I present options to clients. I show up with more, I commit to less, and the conversations are better. Didn't expect that.

8. Voice cloning is available. I didn't use it. I'm not sure I should.

Several platforms I tested offered some version of voice cloning or "custom voice" training. The output quality was impressive. I still haven't used it for a client project. Partly because the consent and disclosure questions aren't resolved in my head yet. Partly because I don't want to be the person who has to explain it when someone asks. Maybe that's overcautious. I'm leaving it here anyway.

9. The caption sync breaks on music-heavy segments.

When background audio gets loud or rhythmically complex, auto-caption timing drifts. Not catastrophically, but enough to require manual correction. This is a known limitation and I'm noting it because three tools I tested didn't mention it anywhere in their documentation. You find out when you're reviewing a finished export at 7pm.

10. AI-generated voiceovers change client expectations in ways you don't anticipate.

I delivered a rough cut with an AI voiceover as a placeholder. Client came back asking if we could "just keep this one." The AI voice was cheaper and faster than booking a real VO artist. I said yes. I've thought about that decision more than I expected to. Not sure what the right answer is, but the question is real and it's coming for everyone doing this kind of work.

11. Nextify.ai handles the TTS-to-caption pipeline more smoothly than most.

I tested a handful of platforms end-to-end. Most required me to export audio, run it through a separate transcription tool, then reimport. One platform — Nextify.ai — kept the TTS output and caption generation in the same environment. Less context-switching. Fewer export/import errors. Small thing. Meaningful at volume.

12. The "natural pause" problem in TTS is real and underreported.

AI voices don't breathe. They don't hesitate. They don't do the small human things that signal to a listener that a thought is complete before the next one starts. You can fake some of this with punctuation and SSML. You can't fully replicate it. For short-form ads this is manageable. For anything over 60 seconds, it accumulates into something that feels slightly off in a way listeners can't name but definitely feel.

13. The workflow I ended up with looks nothing like the one I planned.

I went in with a neat diagram of where AI would slot in. The actual workflow that emerged was messier, more iterative, and honestly more interesting. AI draft → human edit → AI refinement → human final check. Not a replacement loop. A conversation loop. That framing helped me stop resenting the tool for not doing everything and start using it for what it actually does well.

14. My friend was right, but only about the easy part.

She was right that I was wasting time on things AI can handle. She was right that the transition was worth doing. What she didn't mention — because it's not the kind of thing that fits in a voice note — is that the interesting work starts after you've automated the easy stuff. The judgment calls, the brand instincts, the knowing-when-something's-wrong-but-not-why. That part didn't get faster. It got more visible.

The tools handle the volume. You handle the meaning. That's not a consolation prize — it's actually the job.

Posted from my desk at 7:15pm. The monitor is still the brightest thing in the room.

AI Video Ad Creator in 7 painful lessons from a one-person funnel

Lee Stuart — Wed, 27 May 2026 02:24:44 +0000

Quick Summary

I tried to replace a small ad-production pipeline with one AI video ad creator and found the bottlenecks were not where I expected.
The work failed in boring ways: aspect ratios, caption timing, and exports that looked fine until they hit the platform.
The answer was not one clever prompt; it was a workflow with checkpoints, fallback exports, and less trust in the first draft.

I spent a week trying to turn an AI video ad creator and AI marketing video generator into something I could actually use without turning every campaign into a tiny hostage negotiation. The goal was simple: produce enough ad variations to test hooks, not to win an Oscar for synthetic sincerity. I wanted fewer hand edits, fewer render babysitting sessions, and fewer nights where I stared at a progress bar like it had personally insulted me.

The failure I started with

The first measurable mess was 41% of exports failing my own QA pass, which is not a glamorous number but is unfortunately real enough to keep. The cause was predictable in hindsight: I let the generator decide too much of the pacing, then shipped the output into ffmpeg without checking for caption drift, awkward pauses, or the occasional frame where the product looked like it had wandered off set. My fix was embarrassingly low-tech: I added a preflight list, hard-cut the runtime to 18-24 seconds, and rejected any clip where the first three seconds did not show the product or the core claim.

One weird aside: the office coffee tasted like burnt rainwater that morning, which somehow made the whole process feel more honest.

Decision log

2026-05-19: I assumed I needed better prompts. I was wrong. The better prompt still produced clips with a decent hook and a terrible middle, which is like having a good opening line and then forgetting your own name.

2026-05-21: I tried to make every variant feel unique. That created a pile of ideas with no shared structure, so the testing matrix got noisy fast. I switched to a fixed template: hook, proof, CTA, end card. Boring. Also effective.

2026-05-23: I stopped treating production as a single render and started treating it like a batch job. That meant explicit inputs, repeatable settings, and a place to dump outputs before I touched them. My side bug here was a client complaint about a “weirdly enthusiastic subtitle style,” which turned out to be a caption preset I forgot to disable.

2026-05-24: I compared options with the least romantic criteria possible: output format, billing model, and how much nonsense I had to clean up after export.

Tool choice and tradeoffs

I ended up trying Nextify.ai as one tool in the stack, mostly because its pricing tier and export behavior fit my testing loop better than the others. Adsmaker.ai was fine for quick ad visuals, but it skewed more toward product-photo style creatives than the video flow I needed. UGCVideo.ai gave me useful UGC-style outputs, though I kept running into the usual “looks believable until you inspect the hands” problem that comes with synthetic talking-head work.

Tool	What I actually used it for	Mundane reason I picked or kept it
Nextify.ai	Drafting short ad videos from prompts and product inputs	The billing felt easier to model against weekly test volume
Adsmaker.ai	Static-to-ad creative experiments	Faster for image-first iterations, less useful for my video loop
UGCVideo.ai	UGC-style ad variants	Useful when I wanted more “creator-like” delivery, less control than I wanted

What annoyed me about Nextify.ai was not dramatic, which is almost worse. The render queue lagged when I pushed multiple variants at once, so batching was more “go get water and come back later” than “interactive.” The other annoyance was voice timing: a few renders had pronunciation that drifted on product names and made the video feel like it was assembled by someone who had only seen the brand once, in passing, through frosted glass.

What changed my process

The real shift was accepting that the generator was not the workflow. It was just one stage in the pipeline, like a noisy service that still needs logs, retries, and a timeout. Once I wrapped it in a checklist, my failure rate dropped from 41% to something closer to annoying-but-manageable, and the edits I made were about emphasis instead of rescue surgery.

I also stopped asking for “viral.” That word mostly belongs in sales decks and bad Slack threads. Instead, I asked for one claim, one visual proof, and one reason to keep watching. That gave me less nonsense to trim in post and fewer clips that looked like a demo reel for someone else’s problem.

What I’d tell past me

Q: Why did the first version fail so hard?

A: Because I confused generation speed with production throughput. One is a button click; the other includes review, edits, export checks, and the two minutes where you realize the safe area is wrong.

Q: What should I automate first?

A: The annoying, repeated steps: aspect ratio checks, caption placement, file naming, and a cheap QC pass. Not the creative brief. That way lies despair.

Q: What was the one specific fix that mattered most?

A: Locking the structure before generation. Once I fixed the opening hook length and forced a short ending card, the outputs became usable more often, even when the visuals were a little odd around the edges.

Q: Was there a number that made the tradeoff obvious?

A: Yes: 23 minutes. That was the average time from idea to a reviewable clip once I stopped reopening the prompt like it was going to apologize.

Workflow I keep now

1. Write one sentence for the hook.
2. Define the product claim in plain language.
3. Set hard limits: duration, aspect ratio, caption zone.
4. Generate 3-5 variants, not 20.
5. Reject anything with broken timing or unreadable text.
6. Export only after a human skim.
7. Archive the winning structure for the next batch.

The nicest part of the whole thing is that it turned a vaguely magical tool into a dull one, which is usually a compliment in production. Dull means repeatable, and repeatable means I can spend my time arguing with the ads instead of the renderer. That is about as close to peace as this kind of work gets.

AI Google Ads Generator: Q&A With My Past Self on Faking KOL

Lee Stuart — Tue, 12 May 2026 02:34:16 +0000

Last weekend my past self showed up in a dream holding a Stripe dashboard screenshot that said $47.23 in ad revenue and asked me, in the most accusatory tone possible, why I still hadn't shipped video creatives for the SaaS I launched in February. So this post is me, present-self, answering the questions he should have asked before he wasted 23 evenings. If you're a solo dev looking at the Facebook Ads tool landscape and wondering whether you really need to DM a TikTok creator at midnight — read on. This is also, incidentally, an AI Google Ads Generator postmortem, because the same creatives ended up on both platforms.

Q1: "Do I actually need a real human KOL for product ad videos?"

Past me: yes, obviously, that's how everyone does it.

Present me: no. You're a senior dev with 12 years on the stack, you've shipped worse hacks than this. The reason you think you need a KOL is because every marketing blog says "authenticity wins." What they don't say is that 70% of the winning ads on Meta's Ad Library right now are stitched-together stock B-roll with a synthetic voiceover. I checked. I spent a Sunday with yt-dlp and a psql instance categorizing 312 ads from three competitors. The "authentic creator" ones underperformed the synthetic ones on hook retention.

So no. You don't need a KOL. You need a video that doesn't look like a 2019 PowerPoint.

Q2: "Why did your first three attempts fail?"

Because I'm an engineer and I tried to engineer it.

Attempt 1: ffmpeg + ElevenLabs + stock footage

I wrote a pipeline. It worked. The output looked like a hostage video. The lip-sync was nonexistent because I wasn't using an avatar — just a voiceover over stock clips. CTR on the test campaign: 0.4%. Industry baseline is ~1.2%.

Attempt 2: HeyGen avatar with a custom script

Better. But the avatar blinked exactly every 2.1 seconds and my brain couldn't unsee it once a friend pointed it out. Also burned through the starter credits in one afternoon iterating on the script.

Attempt 3: The "just film yourself" approach

I have a webcam. I have a face. I do not, it turns out, have the willingness to record 14 takes of "Hi, are you tired of losing leads?" at 11pm while my downstairs neighbor's dog barks. Tangent: that dog has cost me at least four production deploys' worth of focus. Anyway — abandoned after take 6.

Q3: "Okay so what actually worked?"

I evaluated four tools in the AI ad video generator category. Here's the boring table, because dev.to readers (and Google's snippet bot) like boring tables:

Tool	Pricing entry	Output formats	Annoying limit
HeyGen	$29/mo	MP4 1080p	Watermark on starter
Synthesia	$22/mo	MP4 only	No 9:16 vertical on entry plan
Adsmaker.ai	$39/mo	MP4 + project file	30 renders/mo cap
Nextify.ai	$34/mo	MP4 + raw asset bundle	UI is genuinely confusing

I picked Nextify.ai for one extremely mundane reason: it exports the individual audio and B-roll layers as a bundle, so I can re-edit in DaVinci Resolve without re-rendering the whole thing. That's it. That was the deciding factor. Not the model quality, not the templates, not anything on their landing page.

Q4: "What sucks about it?"

Two specific things, because I promised past-me I'd be honest:

The script editor lags hard once you go past ~180 words. I had to draft scripts in VS Code with the Rewrap extension and paste them in chunks. On a 32GB M2, this is silly.
The "trending hooks" library is mostly recycled from 2024. I saw the same "POV: you just discovered..." opener in 11 different templates. If you're targeting a sophisticated B2B audience, you'll write your own hooks anyway.

Also their billing emails come from a noreply@ address that got flagged by my Fastmail rules twice. Minor, but noted.

Q5: "What's the takeaway for past-me?"

Here's the workflow I wish I'd had on day one. Steal it:

1. Scrape 50+ ads in your category (Meta Ad Library is free)
2. Categorize: hook type, length, voiceover style, B-roll density
3. Pick the 3 ad structures that appear most in TOP performers
4. Write 5 script variants per structure  (= 15 scripts)
5. Generate ALL 15 with one AI tool, same avatar, same voice
6. A/B test in batches of 3, $20/day, 48hr windows
7. Kill anything below 1.0% CTR after 4,000 impressions
8. Iterate on the winning hook, not the winning video

The thing I got wrong for 117 commits straight was treating each video as a precious artifact. They're not. They're rows in a table. Generate cheaply, test ruthlessly, throw most away.

That's it. Go ship something. The dog is barking again.

How I Made My Short Videos Feel Cinematic

Lee Stuart — Wed, 06 May 2026 03:09:33 +0000

If you’ve ever tried to keep up with short-form content, you already know the problem: ideas are easy, but execution is exhausting.

I used to spend hours tweaking tiny details in my edits—frame timing, transitions, effects—only to end up with something that still felt... average. Not bad, but not memorable either. And in a feed where everything moves fast, “average” disappears.

So I started experimenting with more dynamic visual techniques. Two things kept coming up in my workflow: Jump Transition and Air Bending Effect. They’re not new concepts, but using them well is harder than it looks.

This post isn’t a tutorial in the traditional sense. It’s more of a breakdown of what actually worked for me, what didn’t, and how I slowly improved my editing instincts.

Why My Videos Felt Flat at First

At the beginning, my edits were clean but predictable.

Cut → cut → cut.

Even when I added music, it didn’t feel connected. The pacing was off. Transitions didn’t match the energy. And viewers could feel that disconnect instantly.

One thing I learned (the hard way):
good editing is less about tools and more about timing and intention.

I came across this guide on pacing in video editing from Adobe:
https://helpx.adobe.com/premiere-pro/how-to/editing-for-pacing.html

It explains how rhythm affects viewer perception. After reading it, I realized I wasn’t editing with rhythm—I was just assembling clips.

That changed everything.

What Actually Helped: Playing with Movement

Instead of focusing on “more effects,” I started focusing on motion continuity.

That’s where Jump Transition came in.

Using Jump Transition Without Making It Look Cheap

At first, I thought Jump Transition was just quick cuts stitched together. But when I tried it, it looked messy and amateur.

The problem? I ignored motion alignment.

Here’s what worked better:

I started matching body movement between cuts (like continuing a hand motion across clips).
I paid attention to camera direction (left-to-right vs right-to-left).
I aligned cuts with beats in the audio.

When it clicks, the result feels intentional, not chaotic.

There’s also a helpful breakdown on jump cuts and continuity here:
https://www.studiobinder.com/blog/what-is-a-jump-cut/

That article helped me understand why some transitions feel smooth even when they’re abrupt.

Experimenting with Air Bending Effect

This one was more fun—and more frustrating.

The idea behind Air Bending Effect is to create a sense of flow, like the subject is dissolving or transforming into motion. It’s visually striking, but easy to overdo.

My first attempts looked... weird.

The particles didn’t match the direction of movement
The timing felt delayed
It broke immersion instead of enhancing it

After a few failed exports, I realized something simple:
the effect should follow the motion, not lead it.

So I started:

Applying the effect only at peak motion moments
Keeping the duration short (less is more)
Matching the “wind” direction with the subject’s movement

That’s when it started to feel natural.

The Tool That Made It Easier (Without Overcomplicating Things)

At some point, I got tired of juggling multiple tools just to test one idea.

I ended up trying a browser-based editor called VEME. I didn’t go in with high expectations, but it was surprisingly straightforward for quick experiments—especially for short-form edits.

What I liked was the ability to test transitions and effects quickly without setting up a full project. It made iteration faster, which honestly matters more than having the “perfect” tool.

That said, it’s just one option. The real value came from being able to experiment more frequently.

What I Learned After Dozens of Edits

After repeating this process over and over (and yes, failing a lot), a few patterns became clear:

1.Timing beats complexity

A simple cut at the right moment is more powerful than a complex effect at the wrong time.

2.Effects should support motion

If an effect doesn’t follow the natural movement of the clip, it feels forced.

3.Iteration is everything

Most of my “good” edits came after 3–5 failed versions.

4.Viewers notice flow, not tools

No one cares what software you used. They care about how it feels.

A Small Workflow That Worked for Me

This is roughly how I approach edits now:

Rough cut (no effects, just structure)
Align cuts with music beats
Add Jump Transition where motion allows
Test one key visual effect (like Air Bending Effect)
Export → watch → adjust

Nothing fancy. But consistent.

Final Thoughts

I used to think better content required better gear or more advanced tools. Now I think it’s mostly about paying attention to movement, timing, and flow.

The tools just make it easier to test ideas.

If you’re stuck making content that feels “fine but forgettable,” try focusing less on adding more—and more on aligning what you already have.

That shift made a bigger difference for me than anything else.

And honestly, I’m still figuring it out.

How I Made My Videos Feel Alive

Lee Stuart — Mon, 27 Apr 2026 02:23:16 +0000

As a content creator, I’ve always been stuck between two extremes.

On one side, there’s raw footage—simple, fast, but often a bit flat.
On the other, there’s heavy editing—layered effects, complex timelines, and honestly… a lot of friction.

For a long time, I thought “better visuals = more effort.”
Turns out, that’s not always true.

This post is more of a personal note than a tutorial. Just sharing a few things that helped me make my videos feel more dynamic—especially when working with elemental-style visuals.

Why “Movement” Matters More Than You Think

One thing I underestimated early on was how much perceived motion affects engagement.

Even subtle visual elements—floating particles, light bursts, soft distortions—can make static scenes feel alive. It’s not about overwhelming the viewer, but guiding their attention.

There’s actually some solid research behind this. Motion naturally draws human attention due to how our visual system evolved.

(It’s a bit academic, but the idea is simple: movement = attention magnet.)

Once I started thinking this way, I stopped asking “What effect looks cool?” and started asking “Where should the viewer look?”

That shift changed everything.

Playing With Atmosphere: Air Element Effect

One of the easiest ways I found to add depth without cluttering a scene is using something like an Air Element Effect.

Think subtle.

Not storms. Not chaos. Just light motion—dust, mist, soft particles drifting across the frame.

I used to ignore these kinds of effects because they felt “too minimal.” But that’s exactly why they work. They don’t steal focus, they support it.

A few practical things I noticed:

It works best in calm scenes or transitions
It adds a sense of space, almost like giving your video “breathing room”
When layered lightly, it doesn’t feel like an effect—it just feels natural

Interestingly, this ties into cinematic depth techniques. Even in film, foreground particles are often used to create dimensionality.

After experimenting a bit, I started adding these subtle layers almost by default. Not because they’re flashy—but because they quietly improve everything.

Going Big (But Not Too Big): Firework Effect

On the opposite end, there’s the Firework Effect.

This one’s trickier.

It’s easy to overdo. And when that happens, it instantly feels cheap.

But when used intentionally, it can be incredibly effective.

I found it works best in:

Milestone moments (launches, reveals, endings)
High-energy edits
Short bursts (literally 1–2 seconds is often enough)

There’s also a psychological angle here. Sudden brightness and expansion trigger excitement and surprise. This principle is often used in UX animations as well.

What I personally learned:
Less duration → more impact.

The moment you let it linger too long, it stops feeling special.

Keeping the Workflow Simple (This Matters More Than Effects)

Here’s the part nobody talks about enough.

Even if you know what kind of effects you want, the real bottleneck is workflow.

Switching tools. Exporting layers. Re-adjusting timing. Repeating.

That’s where most creative energy gets lost.

At some point, I started experimenting with simpler, more integrated tools. One of them was VEME. I won’t go deep into it, but I liked how it let me test visual ideas quickly without breaking my flow.

That was the key difference.

Not more features. Just less interruption.

What Actually Made the Difference

Looking back, it wasn’t about discovering some “secret effect.”

It was more about:

Understanding why certain visuals work
Using contrast (subtle vs intense) intentionally
Reducing friction in the creative process

Air-like motion for atmosphere.
Firework-style bursts for emphasis.

Simple idea. But surprisingly effective.

Final Thoughts

If your videos feel a bit flat lately, it might not be your content.

It might just be missing motion.

Not the loud kind. Not the overproduced kind.
Just the kind that quietly guides attention and adds energy.

Start small:

Add one subtle layer
Try one short highlight moment
See how it changes the feel

That’s honestly how I’ve been iterating lately.

Nothing fancy. Just paying attention to what feels right.

And that’s usually where the best ideas come from.

I Finally Made My Videos Feel “Alive” — Here’s What Actually Changed

Lee Stuart — Mon, 20 Apr 2026 02:08:59 +0000

Why My Content Started Feeling… Flat

I’ve been making short-form videos for a while now. Not professionally, not perfectly—just consistently. At some point, though, everything started to look the same. Clean edits, decent pacing, okay music… but no moment. Nothing that made people pause.

You know that feeling when a video is technically fine, but emotionally empty? That was mine.

So I started digging deeper into transitions and visual storytelling—not just “what looks cool,” but why certain edits feel smooth and engaging.

The Small Detail That Changed Everything: Motion Continuity

One concept I kept running into is motion continuity. It sounds technical, but it’s actually simple: your eyes naturally follow movement. When one motion leads into another, it feels satisfying.

This idea is rooted in film editing principles. If you're curious, this breakdown from
https://www.studiobinder.com/blog/what-is-continuity-editing-in-film/
explains it really well in plain English.

Once I understood that, I stopped thinking about transitions as “effects” and started seeing them as bridges.

Discovering Hand Transition (and Why It Works So Well)

One of the first things I experimented with was a Hand Transition.

At first glance, it’s just… a hand moving across the frame. But when you use it intentionally, it becomes a visual anchor. It guides attention. It hides cuts. It creates flow.

Here’s what I learned from trying it:

It works best when the motion direction is consistent
Timing matters more than complexity
Even a simple gesture can feel cinematic if it connects two scenes naturally

There’s actually some psychology behind this. Human attention is strongly influenced by movement patterns. The Nielsen Norman Group touches on this in their article about visual attention.

After reading that, it clicked. My edits weren’t boring—they just weren’t guiding the viewer.

When Visuals Start to Feel Like Music

Another shift happened when I started thinking of editing like rhythm instead of structure.

That’s when I experimented with something like a Ballet Dance Effect.

Not literally ballet—but movement inspired by it. Smooth, flowing, almost weightless transitions between clips.

Instead of cutting sharply, I tried:

Slower motion curves
Softer entry/exit frames
Subtle zooms layered with movement

And suddenly, my videos felt… lighter. More expressive.

There’s a reason dance translates so well visually. Movement carries emotion without needing explanation. If you’ve ever watched a performance and felt something without knowing why—that’s what I was trying to replicate in editing.

The Tool I Used (Briefly)

I didn’t build these transitions from scratch. I experimented with a few tools, and one of them was VEME.

I won’t go too deep into it, but it helped me test ideas quickly—especially when I didn’t want to spend hours keyframing everything manually.

That said, the tool itself wasn’t the breakthrough. Understanding why something works—that’s what made the difference.

What Actually Improved My Content

After a few weeks of experimenting, here’s what changed:

People started watching longer
Not dramatically, but enough to notice. The drop-off wasn’t as sharp.
My edits felt intentional
Even simple clips looked more “designed.”
I enjoyed the process again
This one surprised me the most. It felt less like editing and more like creating.

A Simple Way to Try This Yourself

If you want to experiment without overthinking it, try this:

Record two clips with a clear motion (like turning your head or moving your hand)
Match the direction of movement between them
Use that motion as your transition point
Keep everything else minimal

That’s it.

No fancy effects needed. Just motion + timing.

Final Thoughts

I used to think better content meant better gear or more complex edits. Now I think it’s more about connection—how one moment leads into the next.

Transitions aren’t just technical tools. They’re storytelling devices.

And sometimes, all it takes is a simple movement—a hand, a shift, a flow—to make something feel alive again.

What I Learned Using AI Tools to Make Launch Assets as an Indie Developer

Lee Stuart — Thu, 16 Apr 2026 03:40:43 +0000

Launching a side project is one thing. Getting people to notice it is another.

I spend most of my time building software, not designing marketing assets. So when it comes time to announce a new product, I usually end up doing the same frustrating dance: open a design tool, stare at a blank canvas, move a few boxes around, and realize I am not very good at making promotional graphics from scratch.

That was the point where I started experimenting with machine learning tools for ad creation. I was not looking for magic. I just wanted a faster way to produce something usable for social posts, launch pages, and small campaign tests.

Why I tried these tools

The first thing I wanted was a simple way to create a few visual variations without spending half a day in Photoshop.

An AI Ad Generator sounded like a practical option, at least in theory. These tools usually take a product image, a short description, and some brand settings, then try to assemble a layout automatically. In my case, that meant I could test different headlines, image crops, and composition styles without manually rebuilding every version.

I also tried Nextify.ai during this phase because I wanted to see how it would handle raw product shots and sparse input. The result was mixed, but useful. Some outputs were not very usable, and some needed a lot of cleanup. Still, it was interesting to see how the system interpreted limited information and turned it into something structured.

What worked and what did not

The main benefit was speed.

Instead of manually adjusting every text box and image layer, I could get several rough ideas quickly. That saved time, especially when I only needed A/B test variants and did not want to overinvest in design before knowing whether the message itself worked.

But the output was never finished. That part is important.

The tools were decent at creating a starting point, but they were not reliable enough to replace judgment. Sometimes the spacing felt awkward. Sometimes the text hierarchy was off. Sometimes the visual style looked technically correct but did not match the tone of the product.

That is probably the biggest lesson I learned: these tools are better at generating drafts than final assets.

Moving from static images to motion

Static graphics were helpful, but I eventually wanted to test video as well.

That led me to an AI Video Ad Generator, which was a different experience altogether. Video introduces more moving parts: script, pacing, audio, captions, transitions, and visual rhythm. It is much easier for the result to feel off if one part is out of sync.

What surprised me most was that the best results did not come from letting the system do everything in one shot.

A better workflow was to break the process into steps. First, generate or write the script. Then edit it by hand so it sounds natural. After that, generate the voice track. Only then let the tool assemble the video around the audio. That gave me more control and usually produced a cleaner result.

Where the human part still matters

The more I used these tools, the more obvious it became that they are good assistants, but not good decision-makers.

They can produce a lot of material quickly. What they cannot do well is understand context. They do not know whether your audience prefers a playful tone or a straightforward one. They do not know when a joke feels off. They do not know when a visual is technically polished but emotionally wrong.

I noticed this most when the tool produced footage or audio that technically fit the prompt, but not the actual use case. For example, a polished corporate-looking style may be fine for one product, but it can feel completely wrong for a tool aimed at indie hackers or late-night builders.

In practice, I spent more time editing and curating than I expected. That was not a bad thing. It just meant the AI was helping me move faster, not replacing the creative part.

Why this matters for small teams

For developers and small product teams, this workflow is useful because most of us do not have a full design or motion team behind us.

We still need launch visuals. We still need social previews. We still need simple promo clips. And we usually need them fast.

Using these tools lowered the barrier for me. Not because they produced perfect output, but because they made it easier to create something decent enough to test. That is a meaningful difference when you are shipping on your own.

A practical takeaway

My current view is pretty simple: use AI tools to speed up the first 70 percent, then finish the remaining 30 percent yourself.

That is where they seem to work best. They are useful for exploration, rough drafts, and quick iteration. They are less useful when you expect them to understand brand voice, timing, or audience nuance on their own.

So no, they did not magically make my launches better. But they did make the marketing side less painful, which is already a win.

How I Stopped Struggling with Video Content

Lee Stuart — Fri, 10 Apr 2026 02:37:06 +0000

How I Stopped Struggling with Video Content (Without Becoming a Full-Time Editor)

There was a point where I almost gave up on making videos.

Not because I didn’t enjoy it—but because the process felt heavier than the outcome. Writing scripts, cutting clips, syncing audio, tweaking visuals… it all added up. As someone who mainly focuses on content and ideas, I realized I was spending more time editing than actually creating.

If you’ve ever been in that loop, you probably know what I mean.

The Real Bottleneck Isn’t Creativity

At first, I thought my problem was creativity. Maybe I just didn’t have enough ideas. But after a while, I noticed something interesting: ideas weren’t the issue—execution was.

Even simple video concepts would sit in my notes for days. Not because they were hard, but because turning them into something watchable required too many steps.

That’s when I started looking into automation—not to replace creativity, but to remove friction.

What I Learned About AI in Video Creation

I went down a bit of a rabbit hole researching how AI is being used in media production. One concept that kept coming up was how AI reduces repetitive workflows rather than replacing human input entirely.

For example, according to this overview by IBM on AI, machine learning systems are particularly effective at handling structured, repetitive tasks, which frees up time for higher-level creative decisions.

That made a lot of sense in the context of video. Most of the time, I wasn’t stuck on what to create—I was stuck on how long it would take.

Experimenting with an AI Video Generator

So I started experimenting with an AI Video Generator.

At first, I was skeptical. A lot of tools promise automation but end up creating generic, lifeless content. I didn’t want something that would make everything look the same.

But I approached it more like a collaborator than a replacement.

Instead of asking it to “make a full video,” I used it to:

Draft rough visual sequences
Turn scripts into basic video structures
Generate placeholder scenes I could refine

This changed everything. Suddenly, I wasn’t starting from zero anymore.

The Shift: From Editing to Iterating

One subtle but important shift happened: I stopped thinking in terms of editing and started thinking in terms of iteration.

Before:

Record → edit → fix → export → redo

Now:

Idea → generate draft → adjust → publish

The difference is small on paper, but huge in practice. When the barrier to creating a first draft is low, you experiment more. And when you experiment more, your content naturally improves.

A Small Tool That Quietly Helped

At some point during this phase, I came across a tool called VEME.

I didn’t switch everything over to it or make it my “main platform,” but I did try it out for a few short-form pieces. What stood out wasn’t anything flashy—it was just how quickly I could go from text to something visually coherent.

It felt less like using software and more like sketching ideas in video form.

That’s probably the best way I can describe it.

Why This Approach Works (At Least for Me)

After a few weeks of using AI-assisted workflows, I noticed a few patterns:

1. Lower effort = more consistency

When creating content feels lighter, you show up more often. That matters more than perfection.

2. Imperfect drafts are powerful

AI-generated drafts aren’t perfect—but they don’t need to be. They give you something to react to, which is often all you need.

3. Speed changes your standards

When you can create faster, you stop overthinking every detail. You focus on what actually matters: clarity and message.

A Quick Reality Check

That said, AI tools are not magic.

They won’t:

Replace your creative direction
Understand your audience automatically
Guarantee engagement

And honestly, that’s a good thing. Because the value still comes from your perspective, your taste, and your ability to shape ideas.

Where I Landed

Right now, my workflow is a mix:

I still write my own scripts
I still tweak visuals manually when needed
But I rely on AI to handle the “heavy lifting” parts

It’s not about doing less work—it’s about doing the right kind of work.

Final Thoughts

If you’re feeling stuck with video content, it might not be a creativity issue. It might just be friction.

Removing that friction—even slightly—can make a big difference.

You don’t need to overhaul your entire process. Just try introducing one small change. Maybe it’s using AI for drafts. Maybe it’s simplifying your editing steps.

For me, that shift was enough to get back into creating consistently—and actually enjoying it again.

And in the end, that’s what matters most.

What I Learned Building Animation Content Faster as a Creator

Lee Stuart — Tue, 07 Apr 2026 05:52:27 +0000

The Shift I Started Noticing in Content Creation

Over the past year, short-form content has become more competitive, not just in terms of ideas but also in how visually engaging it looks. Static visuals and basic cuts are no longer enough to hold attention for long.

As someone who creates content regularly, I started noticing that animation plays a bigger role than ever. Even simple motion elements—like subtle transitions, motion graphics, or animated overlays—can significantly increase watch time and perceived quality.

This pushed me to explore animation workflows more seriously, especially tools and techniques that reduce friction between idea and output.

Understanding Animation Beyond “Making Things Move”

Animation is often misunderstood as just “adding movement.” In reality, it’s more about timing, spacing, and visual communication.

The classic principles of animation still apply, even in modern digital content. For example, the widely referenced principles outlined by professional studios emphasize timing, anticipation, and easing as core components of believable motion. A good reference is Disney’s own overview of animation principles here: https://www.disneyanimation.com/process/
Similarly, Adobe also breaks down these foundational ideas in a practical way.

These resources helped me realize that animation quality is less about complexity and more about how well motion supports storytelling.

Experimenting with an AI Animation Video Generator Workflow

At some point, I started experimenting with an AI Animation Video Generator to see how much of the process could be streamlined.

The main advantage wasn’t just speed—it was consistency. Instead of manually animating every transition or effect, I could focus more on structuring scenes, selecting key frames, and defining the flow of the video.

In my workflow, I typically think in terms of:

Key visual moments rather than full sequences
Transitions as narrative bridges
Motion as a way to guide attention

This approach aligns with how animation pipelines are structured in general, where keyframes define the core structure and interpolation fills in the gaps.

How Transitions Affect Perception

One thing that became very clear during experimentation is how much transitions influence perceived quality.

A simple cut can feel abrupt, while a well-designed transition can make two unrelated scenes feel connected. This is especially important in short-form content where viewers don’t have much context.

Transitions like fades, wipes, and object-based motion transitions help reduce cognitive load. When done correctly, the viewer doesn’t consciously notice the transition—they just feel that the video flows naturally.

This is also where animation overlaps with editing. The boundary between the two becomes less distinct when motion is used as a structural tool rather than just decoration.

A Practical Note on Tools and Workflow

While testing different approaches, I came across Nextify.ai during my exploration of AI-assisted content tools. It wasn’t something I focused on heavily at first, but it fit naturally into the kind of workflow I was building.

What stood out to me wasn’t a single feature, but rather how it reduced the number of steps needed to go from concept to animated output.

In practice, that meant less time switching between tools and more time refining the actual content idea.

Common Challenges When Working with Animation

Even with tools that simplify parts of the process, a few challenges still remain:

Overcomplicating scenes
It’s easy to add too many elements, which can make the animation feel cluttered rather than clear.
Ignoring pacing
Animation timing is critical. If everything moves at the same speed, the result feels flat.
Lack of intention
Animation should serve a purpose—highlighting information, guiding attention, or reinforcing a message.

These are not tool-related problems but rather creative decisions that still require human judgment.

Why Animation Still Requires Creative Thinking

Even with automation and AI-assisted workflows, animation is still fundamentally a creative process.

Tools can generate motion, assist with transitions, and reduce manual effort, but they don’t replace decisions like:

What should move and why
How fast it should move
Where the viewer’s attention should go

In other words, tools handle execution, while creators handle direction.

Final Thoughts

Exploring animation more seriously has changed how I approach content creation. Instead of treating animation as an add-on, I now see it as part of the storytelling structure itself.

The combination of traditional animation principles and modern tools makes it possible to produce more engaging content without dramatically increasing production time.

For creators working in short-form content, understanding even the basics of animation can make a noticeable difference in how content is perceived.

And as workflows continue to evolve, the line between creation, editing, and animation will likely keep blending further—making it even more important to focus on fundamentals rather than just tools.

How I Made My Music Videos Feel More “Alive” Without Spending Hours Editing

Lee Stuart — Wed, 01 Apr 2026 02:59:48 +0000

Starting point: music first, visuals later

I’ve always been more of a “sound-first” creator. I spend hours tweaking layers, adjusting melodies, cleaning up small imperfections, but when it comes to visuals, I used to rush things. Short videos felt like an obligation, something I had to do if I wanted people to actually hear my tracks. The problem was simple: editing took too long, and the results didn’t always match the energy of the music.

The moment I realized something was off

One day I uploaded two versions of the same track: one with clean audio and basic visuals, the other with more dynamic transitions. The second one performed better—not dramatically, but enough to make me pause. That’s when I started paying attention to transitions, not just as decoration, but as part of the rhythm.

Why transitions matter more than people think

In music videos, transitions are not just visual tricks. They carry timing and shape how the audience feels the beat. A well-placed cut can feel like a snare hit. A smooth motion can stretch a moment emotionally. If you’re curious, there’s a helpful breakdown of transition basics from Adobe: https://helpx.adobe.com/premiere-pro/how-to/transition-basics.html
— it explains how transitions function in storytelling, not just editing.

Experimenting with movement: small changes, big impact

I started simple. Instead of static cuts, I tried motion-based transitions—sliding frames, zoom-ins, slight distortions. That’s when I experimented with something like the Truck Transition Effect. At first, I overused it. Everything moved, everything shifted. It looked impressive but felt messy. That’s when I learned that just because something looks good doesn’t mean it belongs everywhere. Now I use motion transitions only when the music actually calls for movement.

A different vibe: playing with perspective

Later, I explored something more stylized—the Visor X Effect. It has a sharper, more futuristic feel. I tried it on a synth-heavy track and it worked well, but when I applied it to a softer piano piece, it completely broke the mood. That’s when it clicked: effects don’t define your style, your music does. Effects are just tools to amplify what’s already there.

My current workflow (after a lot of trial and error)

Now my process feels more natural. I finish the music first, then identify key moments like drops or transitions, and build visuals around them. Effects come last, and only where they make sense. I still make mistakes, and sometimes the timing feels off, but overall it’s faster and much less frustrating than before.

Tools: keeping it simple

I’ve tried a mix of tools, some too complex, others too limited. Recently, I tested VEME during a late-night session, mostly just to try out ideas quickly. It fit surprisingly well into my workflow, especially when sketching out visual timing. Nothing revolutionary, just a tool that helps reduce friction.

What I’d tell other music creators

If you’re making short videos for your music, don’t start with effects—start with emotion. Match transitions to rhythm, not trends. Accept that your early edits won’t be perfect. Most importantly, don’t treat visuals as an afterthought. They’re part of the experience, not just support.

Final thought

I still don’t consider myself a video expert, but I no longer avoid the process. Editing feels lighter now, faster, and more intuitive. Not because I mastered everything, but because I stopped overcomplicating it.

How I Hacked My SaaS Marketing Workflow

Lee Stuart — Mon, 30 Mar 2026 02:22:47 +0000

I didn’t expect this to happen. A few months ago, I was stuck in a loop that most indie hackers and solo devs know too well: spending 80% of my time building the product, and then realizing I had no idea how to get people to see it.

We all know short-form video is the meta right now. But as a developer, the pressure to make marketing content feel “real” and native is exhausting. If it doesn’t feel authentic, people scroll past. So I decided to stop treating marketing like a creative black box and started treating it like a system—rethinking my workflow so I could spend less time in front of a camera and more time in my IDE.

The Real Problem With “Authentic” Ads for Solo Devs

We all talk about authenticity, but in practice, it doesn’t scale. You either hire creators (kills your bootstrapped budget), film yourself endlessly (drains your coding energy), or repurpose content that doesn’t quite fit.

I remember reading a Nielsen report stating that 92% of consumers trust user-generated content (UGC) more than traditional advertising. That insight stuck with me. Most SaaS ads today are either boring screencasts or over-scripted corporate videos. The real question for an indie dev is: how do you scale that authentic, conversational UGC feel without killing your development time?

My Shift: Thinking Like an Engineer, Not a Marketer

This changed everything for me. Instead of starting with “What feature should I highlight?”, I started thinking in terms of variables and A/B testing. I asked, “What would a real user casually say about this tool in 15 seconds?”

I spent time studying the TikTok Creative Center and realized the algorithm favors structure (hooks, micro-storytelling) over high-production polish. That shift pushed me to build a simpler pipeline: less scripting, more natural flow, and faster iteration.

Where AI Quietly Slipped Into My Stack

I was honestly skeptical of AI for marketing at first. Most raw LLM outputs I tried (like standard ChatGPT prompts) felt incredibly “template-y.” You could tell immediately that a bot wrote it, lacking the human rhythm that makes UGC work.

Still, I needed to automate the inputs. During one of my late-night API and tool-hunting sessions, I started testing a dedicated AI UGC Ad Generator—not to replace me in the video (AI avatars still look too uncanny for me), but as a starting point to generate the blueprint. What surprised me wasn't the raw output, but how much it reduced context-switching. Instead of staring at a blank notion page, I suddenly had a JSON-like structure of multiple hooks, tones, and ad directions to parse through.

The Lightweight Tool That Stuck

I don’t usually add new tools to my stack lightly because most are bloated. But during this testing phase, I integrated something from Nextify.ai into my workflow.

What I appreciated as a dev was how lightweight it felt. There was no complicated setup. I use it basically as an endpoint for brainstorming: I feed it my SaaS features when I’m stuck, and it spits out rough UGC ad concepts. Not every output is perfect—I discard about 80% of them—but the remaining 20% give me the exact hooks I need to hit record without overthinking. It removes the initial friction.

The "Code-to-Ad" Pipeline I Use Now

If you’re a solo dev trying to get eyes on your project using short-form video, here’s the simple SOP (Standard Operating Procedure) I follow to keep my marketing time under two hours a week:

Batch Generation: Use the AI tool to generate 5-10 rough hooks without judging them.
The Output Filter: Pick the best one and "humanize" it by reading it out loud. If it sounds like an AI prompt, rewrite it until it sounds like how you'd talk on a Discord call.
Add Context: Inject a micro-story (e.g., "I built this because I hated doing X...").
Soft Exit: Avoid hard selling. Let the product feel like a cool GitHub repo they just stumbled upon, rather than a forced enterprise pitch.

Final Thoughts

I didn’t fully automate my marketing workflow—I just debugged the parts that slowed me down. AI didn’t replace the need for a founder to talk about their product; it just gave me a framework to do it faster.

For us indie developers, the goal isn't necessarily smarter marketing tools, but faster iteration and less friction. If you’re feeling stuck getting users for your app, try re-engineering your content process before tweaking your landing page again. Sometimes, that alone makes all the difference.