PromptZone - AI Prompts, Guides and Tools for Builders: daniel

I Accidentally Turned a Small Market Into a Festival

daniel — Mon, 31 Aug 2026 08:01:55 +0000

I started with a pretty ordinary event brief.

Saturday morning. A small outdoor market with coffee, baked goods, handmade ceramics, and a musician near the stalls.

That was enough for the idea.

Then I started writing the video prompt.

Somewhere along the way, the small market picked up a sweeping entrance shot, a much bigger crowd, rows of busy stalls, golden light, extra performers, and a dramatic final view.

It sounded good when I read it back.

It just wasn't the same event anymore.

Back to the Original Notes

I deleted most of what I'd added and looked at the brief again.

Coffee being poured made sense.

Someone browsing the ceramics made sense.

The musician could stay in the background.

One wider view could show the outdoor setting.

I didn't need much more than that.

I kept the same notes open when I made a few Wan 3.0 drafts. That made it easier to spot the moment a scene started drifting.

In one version, the market looked much larger and busier than I'd imagined it. There was nothing obviously wrong with the image itself. If I'd seen it without the brief, I probably wouldn't have questioned it.

But it wasn't what I'd started with.

So I went back rather than trying to make the new version fit.

Some Things Can Change. Others Can't.

A fictional pottery workshop gave me a similar problem.

My notes said:

Sunday at 2 PM

Beginner class

Materials included

Eight places available

There's plenty I can invent visually around that.

The tables can be wooden. There can be clay on the workbench. Finished cups might sit on shelves in the background.

But I can't quietly turn Sunday into Saturday.

And if there are eight places, I shouldn't turn that into twenty because a fuller room happens to look better.

The same goes for a real price, address, product name, opening time, or offer.

Those details come from the source, not from whatever looks nicest in the generated scene.

This matters more than I expected because once a draft looks polished, invented details stop feeling invented.

They just look like part of the video.

I Don't Need to Show the Whole Brief

I used to make this harder by trying to turn every line into a shot.

Say a bakery brief mentions fresh bread, takeaway coffee, outdoor seating, opening hours, pastries, and a weekend menu.

That's useful information.

It doesn't all need screen time.

Maybe the video only shows the first part of the morning: the door opens, bread comes out, coffee is poured, and someone leaves with a paper bag.

The menu can stay on the page.

The opening hours can stay in the caption.

That leaves the video with one simple job instead of six.

Sometimes the bigger problem is that the original notes don't give me anything useful to show.

"Great food, friendly atmosphere, perfect weekend spot" sounds fine in copy, but I can't do much with it visually.

"Small bakery counter, cinnamon rolls coming out of the oven, two tables outside, customers picking up coffee" is much easier.

It's not better because it sounds more creative.

It's better because I can see what happens.

I Check the Draft Against the Thing I Started With

When a draft is ready, I still keep the original notes nearby.

I'm looking for obvious visual problems, but I'm also checking for smaller changes I might have stopped noticing.

Maybe the shop has become much bigger.

Maybe there's a service in the scene that the business doesn't offer.

Maybe a quiet local event suddenly looks packed.

Maybe a product appears to do something that was never part of the source material.

With a fictional exercise, I can decide whether any of that matters.

With a real business, event, or product, I don't make that assumption.

The same goes for actions that look convincing on screen. If generated footage appears to show how food is prepared, how a product works, or how a tool is used, I don't treat the generated motion as an instruction or proof that the action is accurate. I go back to the real source before using that kind of detail publicly.

A polished draft can still get the basic facts wrong.

That's why the boring notes stay open on my screen.

I started with a small Saturday market.

I'd like it to still be a small Saturday market when I'm finished.

I Tried to Make a Simple AI Video More Interesting. I Overdid It.

daniel — Fri, 28 Aug 2026 07:05:10 +0000

I wanted a simple shot of a sparkling-water bottle on a café table. Someone would pick it up, open it, take a drink, and put it back down.

By the time I finished the prompt, I'd somehow added a slow camera move, condensation close-ups, people walking behind the table, a lighting change, and another product shot at the end.

Watching it back, I wasn't wondering what else the scene needed.

I was wondering why I'd put so much in it.

I Tried the Same Scene With Less Going On

For the next version, I kept the bottle on the table and let the opening sit for a moment.

A hand came into frame, picked up the bottle, and opened it. The camera moved slightly closer during the action, then stopped.

That was about it.

The second version wasn't as flashy, but it was much easier to watch. I knew where to look, and the main action didn't have to compete with something happening in the background.

It also made me notice how often I was adding instructions simply because I could.

Now I Decide Where the Scene Ends First

This has become a small habit.

Before thinking about camera movement, lighting, or visual style, I decide what I want to see at the end.

For the café scene, I wanted the final moment to come after the drink, with the bottle still visible on the table.

Once I knew that, the rest became easier.

If the last frame is the bottle back on the table, I don't need to invent a second location halfway through the clip.

I started doing this during a few longer Seedance 2.5 generations. Instead of thinking about how much I could fit into the available time, I worked backward from the last moment I actually wanted.

That kept the prompts much more focused.

The Camera Doesn't Have to Be Busy Either

I ran into the same problem with a coffee machine scene.

All I needed was a mug under the machine, someone pressing the button, and coffee pouring into the cup.

My first version also asked the camera to move around the machine.

It wasn't completely wrong. It just made an ordinary action feel more dramatic than it needed to be.

On the next try, I kept the framing fairly steady and used a small move toward the mug when the coffee started pouring.

That suited the scene better.

Now, when I'm tempted to add another camera instruction, I ask whether it helps me see the action more clearly.

If removing it changes nothing important, I usually leave it out.

I Don't Rush Through Every Pause Anymore

Quiet moments used to bother me.

If nothing happened for a second or two, I'd assume the scene needed another action.

But think about something as ordinary as picking up a package from the front door.

You pick it up.

You look at the label.

Then you open it or carry it inside.

That moment spent looking at the label isn't wasted. It's just part of the action.

The same applies at the end of a shot. Sometimes holding on the product for a moment feels more natural than squeezing in another camera move before the clip stops.

I have taken this too far.

In one test, I left a long pause at the beginning because I thought the scene needed more breathing room.

It didn't.

It just felt slow.

I shortened the opening on the next version.

So I'm not trying to make everything slower. I just don't assume that every quiet second needs fixing.

I Stopped Trying to Fill the Full Duration

This was another easy habit to fall into.

If I had room for a longer clip, I'd start looking for more things to put in it.

Now I let the action decide.

Say the idea is a pair of sunglasses in a case.

The case opens.

Someone picks up the glasses.

The shot ends on the side profile.

If that already feels finished, I don't need another location or another action just to make the clip longer.

A different scene might naturally take more time. Someone could enter a room, notice something on a table, walk over, and interact with it.

That's fine too.

I just don't choose the length first and then try to invent enough material to fill it.

One Check Before I Generate

Before generating, I sometimes read the idea again without words like "cinematic," "dynamic," or "beautiful."

If I can still tell what's supposed to happen, the scene is usually clear enough.

If all that's left is a location and a mood, I probably need to think about the action a little more.

There's also one boundary I keep in mind when these clips are meant for public use.

I treat generated scenes as visual drafts, not product evidence. If a scene shows how something works, how food is prepared, or how a tool is used, I check the real instructions and source material separately rather than trusting generated motion.

A convincing-looking clip can still show something that isn't accurate.

When a scene feels crowded now, I usually remove something before adding another instruction.

Sometimes it's a camera move.

Sometimes it's background action.

Sometimes it's an ending I added because the original ending seemed too plain.

And sometimes I cut too much and have to put a little movement back.

But I've stopped treating every quiet second as unfinished.

If I can clearly see one small thing happen from start to finish, I usually leave it there.

I Was Writing My AI Video Prompts Backwards

daniel — Mon, 17 Aug 2026 06:23:45 +0000

For a while, my AI video prompts kept getting longer.

If I wanted to make a simple product clip, I'd describe the product, room, lighting, camera angle, movement, mood, colors and probably five other things I didn't really need.

Sometimes it worked.

Other times I'd get a nice-looking video that had very little to do with what I had pictured.

Eventually I realized I was trying to make the prompt do everything.

So I tried the opposite.

I started showing the model more and describing less.

A Coffee Video Made This Obvious

I was making a short concept video for a fictional coffee brand.

Nothing complicated. Just a package on a table, warm morning light and a slow camera move toward the product.

My first prompt was way too long.

I described the wooden table, the café, the packaging, the sunlight, the shadows, the lens and the camera movement.

The result looked polished, but the mood was wrong.

So I started again.

This time I used a product image for the package, a café photo for the lighting, and a short clip that had roughly the camera movement I wanted.

Then the written instruction became:

Keep the product as the focus. Match the warm morning light from the café reference and follow the slow forward movement from the video reference.

Much simpler.

More importantly, I knew what each input was supposed to control.

I Started Giving Every Reference One Job

That became my new rule.

If I add a reference, I should be able to explain why it's there in a few words.

For example:

Image 1 → product appearance
Image 2 → lighting and mood
Video 1 → camera movement
Audio 1 → rhythm and atmosphere

This clicked for me while testing reference-heavy prompts with the Seedance AI Video Generator. Instead of asking one giant prompt to carry the whole idea, I could let different references handle different parts of it.

It also made troubleshooting much easier.

If the movement was wrong, I knew where to look.

If the mood was wrong, I didn't immediately rewrite the whole prompt.

More References Can Actually Make Things Worse

I learned this one by making a mess.

For another product test, I used a bright studio product photo, a dark café interior, and a fast handheld camera reference.

Individually, I liked all three.

Together, they made very little sense.

The result wasn't exactly bad. It just looked like it couldn't decide what kind of video it wanted to be.

That's when I stopped treating reference slots like a checklist.

Just because I can add another image doesn't mean I should.

Now I would rather use four references with obvious jobs than fifteen references pointing in different directions.

My Prompts Got Shorter After That

This was probably the most unexpected part.

Once the references were doing the visual work, I didn't need paragraphs of description anymore.

Instead of writing:

A beautifully designed premium coffee package sitting on a rustic wooden table inside a cozy European café with warm golden morning sunlight...

I could write something closer to:

Use @Image1 for the product. Match the warm light of @Image2. Follow the slow push-in from @Video1. Keep the package centered and finish on a clean close-up.

It isn't a prettier prompt.

It's an instruction.

That distinction has been useful.

The Four Questions I Ask Before Generating

I don't have a complicated prompt formula anymore.

Before I generate anything, I just check four things:

What absolutely needs to stay the same?
What should actually move?
Do I already have a reference that shows what I mean?
What still needs to be explained in words?

A sneaker video is a good example.

If I already have a clean product image, I don't need half the prompt describing the shoe.

I'd rather spend those words explaining what happens:

The sneaker lands softly on the platform. Camera makes a slow half-orbit and settles on the side profile. Hold the final frame for two seconds.

The image handles appearance.

The prompt handles behavior.

That division makes sense to me.

This Works Beyond Product Videos

I've ended up using the same idea for completely different projects.

For a food clip, one image can establish how the dish should look while a video reference handles the camera move.

For a travel scene, an environment image can establish the location while the prompt focuses on what happens there.

For a storyboard test, I care more about blocking and movement than tiny visual details.

The references change.

The principle doesn't.

Show what is easier to show. Describe what needs to happen.

What I Do Differently Now

My old process was roughly:

Idea → huge prompt → generate → something is wrong → rewrite everything → generate again

Now it's more like:

Idea → choose references → give each one a job → write the missing instructions → generate → fix one variable

It's not a dramatic new technique.

It just removes a lot of guessing.

And when something goes wrong, I have a better idea why.

Final Thoughts

I still enjoy experimenting with prompts.

I just don't think a prompt needs to contain the entire video anymore.

If I already have a picture that shows the exact lighting I want, I'll use it.

If a five-second clip shows the camera movement better than I can describe it, I'll use that too.

The biggest lesson for me has been surprisingly simple:

Prompts are good at telling the model what to do. References are good at showing it what you mean.

Once I stopped asking one giant prompt to do both jobs, making AI videos became a lot less frustrating.

How I Use LTX 2.3 to Present Interior Design Ideas Before They’re Built

daniel — Thu, 18 Jun 2026 06:46:23 +0000

One challenge I face when working with clients is helping them imagine a space that doesn't exist yet.

Floor plans are useful.

3D renders are helpful.

But many clients still struggle to picture how a room will actually feel.

I wanted a faster way to communicate design ideas without producing a full architectural animation.

That's when I started experimenting with AI video workflows.

What

Recently I've been using LTX 2.3 to create short walkthrough-style concept videos from interior renderings.

Instead of replacing professional visualization software, I use it to explore ideas before final presentations.

Some capabilities that fit this workflow include:

Image-to-video generation
Camera path generation
Scene refinement
Multi-modal inputs
Clip extension

The goal isn't photorealistic perfection.

It's helping clients understand the atmosphere of a space.

How

Step 1: Prepare One Render

I usually start with a single image from SketchUp, Blender, or another rendering tool.

It doesn't need to be perfect.

It just needs to communicate the layout.

Step 2: Describe Camera Movement

Instead of describing furniture again, I focus on how the viewer should experience the room.

Example:

slow walkthrough into a modern living room, natural morning sunlight, gentle camera movement, warm wooden textures, architectural presentation style

Step 3: Generate Different Walkthroughs

I normally test:

entrance perspective
eye-level walkthrough
kitchen transition
window reveal

Each version highlights a different design decision.

Step 4: Collect Client Feedback

Instead of sending only static images, I send several short clips.

Clients often give more specific feedback after seeing movement through the space.

Example Prompt

architectural walkthrough, Scandinavian living room, floor-to-ceiling windows, warm oak flooring, soft daylight, slow cinematic camera movement, realistic interior atmosphere

Use Case

This workflow has been useful for:

Interior Designers

Showing room concepts before construction.

Architects

Presenting early design ideas.

Real Estate Developers

Previewing apartment layouts.

Furniture Designers

Visualizing products inside realistic spaces.

Home Renovation Studios

Helping clients compare different design directions.

Why

After using this workflow on several projects, I noticed a few practical benefits:

Faster design presentations
Easier client communication
More engaging concept reviews
Better visualization of spatial flow
Less time producing complex animations

The biggest benefit wasn't rendering quality.

It was helping clients make decisions earlier in the design process.

A short walkthrough often communicates scale, lighting, and atmosphere much better than a single still image.

Final Thoughts

I still create high-quality renders for final presentations.

But during the concept stage, LTX 2.3 has become a useful way to test ideas and explain spaces before investing time in full animation.

If you're working on architecture, interior design, or renovation projects, try turning one render into a short walkthrough instead of sending only static images.

Sometimes a few seconds of movement explain a design better than ten still images.

How I Created Product Videos Without Hiring a Video Team

daniel — Tue, 02 Jun 2026 02:00:11 +0000

I run a small online store, and one challenge I kept running into was product videos.

Professional product videos look great, but for a small business, every new product means:

shooting footage
editing clips
arranging lighting
paying freelancers

Recently I started experimenting with Wan 2.7 AI Video Generator, and it completely changed my workflow.

My Simple Workflow

Step 1: Create a Product Image

I start with a clean product image.

Nothing fancy.

Just a clear product photo with a simple background.

Step 2: Write a Motion Prompt

Instead of describing the product again, I focus on movement.

For example:

Premium smartwatch on a black reflective surface, slow orbit camera movement, soft studio lighting, luxury commercial atmosphere, smooth cinematic motion.

One thing I learned quickly is that motion and camera direction matter much more than adding dozens of visual keywords. Many Wan 2.7 prompting guides recommend focusing on subject, action, camera movement, and pacing instead of keyword-heavy prompts.

Step 3: Generate Multiple Variations

I usually create:

product close-up
rotating showcase
dramatic hero shot
vertical social-media version

This gives me enough material for ads and landing pages.

What Worked Best

After a few days of testing, I noticed:

shorter prompts worked better
one camera movement per shot produced cleaner results
consistent lighting improved brand identity
image-to-video workflows were easier than starting from text alone

Wan 2.7 is designed around generation, editing, references, continuation, and creative control rather than simple one-click video generation.

Example Prompt

Here's one of the prompts that produced surprisingly good results:

Premium coffee machine on marble countertop, slow push-in camera, warm morning sunlight, realistic reflections, commercial advertising style, smooth motion.

Final Thoughts

I'm not replacing professional production completely.

But for:

product launches
social ads
Shopify stores
landing page videos
marketing experiments

Wan 2.7 has become one of the fastest ways I've found to turn a static product image into a usable promotional video.

Curious how other people are using AI video tools for e-commerce workflows lately.