Building a photo processing tool for dealers. Looking for feedback and beta testers

[datepast]

Hey everyone! Glad to be here, been lurking for a bit and this place is genuinely great, tons of real hands-on experience in the discussions.

To introduce myself, I'm software developer and I building a tool for small/mid auto dealers to help with photo processing (the boring repetitive stuff that just takes way too much time).

So, to get to the point, I've been build something that, I hope, could help dealers (or really anyone moving a decent volume of cars a month) save some time on the photo editing. The idea is to get your photos ready for listings on sales platforms (fb marketplace, cargurus, or your own site) and make it more consistent and professional look. Ideally it should be as painless as possible: upload your photos and just forget about it, then download ready images to post.

The app works with photos you already have. Right now it does four things: replaces the background with a clean studio look, hides license plates, generates a vehicle ad description, and makes it easy to add your dealership logo. But one more thing worth to mentioning, it uses AI, but it doesn't touch the car itself. So car stays as in original photos and no misleading potential buyer.

So if anyone's up for giving it a try, it'd be great if you could test it against your actual inventory. Run it through your real workflow of posting a car and see how it fits in. App have a built in paywall with 3 free images after registration, but I'll add everyone free access, so no blockers here.

What I'm really hoping to get out of this:

• Honest feedback on how the tool actually performs on real inventory photos
• A better understanding of the day-to-day problems dealers run into that I might not even be aware of
• Ideas on what else could make this more useful for your workflow
• A sense of whether the output quality is actually good enough to use in real listings
• Any deal breakers - things that would stop you from using something like this even if it worked perfectly

If someone interested, just left adress here carflowai.com. Currently it's available as a desktop web app and on iOS.

ps. left here your mail or dm me, so I can add more credits for free access

pss. I'll reply to every comment, welcome to the discussion!

psss. I Attach couple examples of the app output

 

[ollieheart]

Hey everyone! Glad to be here, been lurking for a bit and this place is genuinely great, tons of real hands-on experience in the discussions.

To introduce myself, I'm software developer and I building a tool for small/mid auto dealers to help with photo processing (the boring repetitive stuff that just takes way too much time).

So, to get to the point, I've been build something that, I hope, could help dealers (or really anyone moving a decent volume of cars a month) save some time on the photo editing. The idea is to get your photos ready for listings on sales platforms (fb marketplace, cargurus, or your own site) and make it more consistent and professional look. Ideally it should be as painless as possible: upload your photos and just forget about it, then download ready images to post.

The app works with photos you already have. Right now it does four things: replaces the background with a clean studio look, hides license plates, generates a vehicle ad description, and makes it easy to add your dealership logo. But one more thing worth to mentioning, it uses AI, but it doesn't touch the car itself. So car stays as in original photos and no misleading potential buyer.

So if anyone's up for giving it a try, it'd be great if you could test it against your actual inventory. Run it through your real workflow of posting a car and see how it fits in. App have a built in paywall with 3 free images after registration, but I'll add everyone free access, so no blockers here.

What I'm really hoping to get out of this:

• Honest feedback on how the tool actually performs on real inventory photos
• A better understanding of the day-to-day problems dealers run into that I might not even be aware of
• Ideas on what else could make this more useful for your workflow
• A sense of whether the output quality is actually good enough to use in real listings
• Any deal breakers - things that would stop you from using something like this even if it worked perfectly

If someone interested, just left adress here carflowai.com. Currently it's available as a desktop web app and on iOS.

ps. left here your mail or dm me, so I can add more credits for free access

pss. I'll reply to every comment, welcome to the discussion!

psss. I Attach couple examples of the app output

I’ve seen something similar with tools like Spyne… the whole “background + logo + description in one flow” idea definitely clicks, especially for busy inventory days.

What stood out to me though wasn’t just the feature set, it was speed. If processing takes too long, it just gets pushed for later. But once it’s quick enough to fit into the daily routine, dealers actually stick with it. Curious what your turnaround time looks like right now?
Also, how are you handling consistency? That’s usually where things get tricky… cars shot in different lighting, weather, random lot spots, and suddenly the listings feel all over the place. It’s easy to make a few cars look good, but scaling that across a full lot is a different story.

 

[Landing Pages Garage]

Hi there,

Your app has a lot of potential, but I believe the current presentation could be much stronger to better reflect its value. Here are my recommendations to help it convert:

• The headline should speak directly to the client’s needs. Instead of a generic title, try something like: "Revamp your car images with one click." Follow this with a benefit-driven subheadline: "Instantly remove backgrounds, blur tags, generate descriptions, and brand your inventory."
• Since the four main subtitles aren't clickable, they feel static. Add a short descriptive line under each to explain the value they offer.
• The current car gallery is a bit repetitive. I suggest using fewer cars but choosing vibrant colors like red, blue, or yellow to catch the eye. Use a title like, "See what our clients are creating."
• Rather than several small clips, create one high-impact, sped-up demo that showcases the entire workflow in seconds.
• I recommend removing the app promo. Since the site is mobile-optimized, a dedicated app is an unnecessary barrier for the user.
• The section with the Mercedes is currently too cluttered with competing fonts and colors. Let the car be the hero—simplify the style to a unified black aesthetic and swap the car for a high-impact red Mercedes or a Porsche to grab attention.
• Build trust immediately by showing your pricing tiers upfront.

I hope this helps! Feel free to send over an update if you’d like more feedback.
Thank you
claudio

 

[levin233]

Honestly, plate hiding + clean background alone would save a ton of time for smaller dealers. Half the battle is just making listings look consistent across Facebook Marketplace and dealer sites. If the car itself stays untouched, that’s actually a big trust factor.

 

[ArmanM]

Lots of background removal tools on the market, I think if I were building one I would build one that utilizes AI to add in banners / key features that is pulled form vin data.

Currently I am using carcutter for our background removal / 360s and CANVA to manually add in banners but if it was auto populated from vin data it would save me a few minutes per vehicle.

 

[datepast]

I’ve seen something similar with tools like Spyne… the whole “background + logo + description in one flow” idea definitely clicks, especially for busy inventory days.

What stood out to me though wasn’t just the feature set, it was speed. If processing takes too long, it just gets pushed for later. But once it’s quick enough to fit into the daily routine, dealers actually stick with it. Curious what your turnaround time looks like right now?
Also, how are you handling consistency? That’s usually where things get tricky… cars shot in different lighting, weather, random lot spots, and suddenly the listings feel all over the place. It’s easy to make a few cars look good, but scaling that across a full lot is a different story.

Hey, great questions!

About speed question, I'll be honest, that's not our strong suit. Processing takes a bit longer than a simple AI overlay because we're doing more under the hood: realistic shadow generation based on the car's actual geometry, not a generic drop shadow. So it's a trade-off, slower, but the result looks more natural.

About consistency question, this is actually one of the core things that separates CarFlow AI from most tools out there. The car itself stays completely untouched. We don't apply any AI overlay on top of it, we cut it out and place it on a new background. That said, I'll be transparent about the edge cases: the cutout follows the car's contour, so very occasionally something at the edge might get clipped, and we do remove what's visible through the windows. So it's not magic, but the car stays exactly as it is in real life, which matters a lot. Misleading buyers with an AI-altered car is the last thing any dealer needs.

Also, big update just shipped a couple of days ago with major improvements to consistency and background quality. Worth checking out at carflowai.com if you're curious

if You want to try it out and give feedback, please let me know, Ill setup free credits on Your account

With best regards

 

[datepast]

Also wanted to share something I'm planning to build next.

Since the logic and service architecture is already there under the hood, I'm planning to spin up a standalone page on carflowai.com where anyone can blur license plates. Completely free, no account needed, just upload and download.

Would anyone here actually find that useful? Curious if there's real demand before I prioritize it