The reason we built Auto Blemish without asking you to pick a brush
Three months before Cleanr launched, a beta tester sent us a message that stuck. 'I love the app, but I'm never going to use object removal because I don't want to spend ten minutes clicking around my own face.' That comment arrived on a Tuesday. By Thursday, we'd started building Auto Blemish.
What blemish removal looked like before we rethought it
Most photo editors ask you to do something odd: zoom in on your own skin, pick a brush size, and manually tap or drag across spots you want gone. It's technical. It requires patience. And if you're editing on a phone after a long day, it's the kind of friction that makes you close the app and accept the photo as is. We'd already built object removal into Cleanr using PatchMatch content-aware technology, which works brilliantly for removing unwanted items from backgrounds. But skin blemishes are different. They're intimate. They're personal. And they shouldn't require you to become a retouching technician. The insight came from watching how people actually use photo editors. They want speed. They want one-tap results. They want to fix a photo in 30 seconds, not 30 minutes.Why Vision face detection changes the equation
Apple's Vision framework does something elegant. It locates faces in an image, identifies facial features, and creates a mask around skin regions. For us, that meant we could say to Cleanr: find the face, map the skin, and then apply blemish removal intelligently without asking the user to do any of the work. Auto Blemish works by analysing the facial regions that Vision detects, then intelligently smoothing and blending skin tone variations, acne, marks, and minor imperfections. No brush. No manual selection. You hit the button, it runs, you move on. That might sound simple, but the engineering matters. We had to balance removal strength so it looked natural, not plastic. We had to handle different skin tones correctly. We had to make sure it didn't over-process and turn your face into a blur. The first version was too aggressive. The second was too subtle. The third one shipped.The difference between Auto Blemish and Face Retouch
Here's a question we get regularly: what's the difference between Auto Blemish (free on Cleanr) and Face Retouch (part of AI Pro)? Auto Blemish handles blemishes, minor marks, texture, and skin imperfections. It's designed to look natural. You won't notice it's been retouched if it's done well. Face Retouch is more comprehensive. It uses Vision face detection to map facial structure and applies selective brightening, smoothing, and enhancement to specific zones (eyes, lips, cheekbones) while leaving overall character intact. Think of it this way: Auto Blemish is a quick fix. Face Retouch is a full portrait session. If you're posting a casual photo to Instagram, Auto Blemish takes seconds. If you're a content creator or small business owner preparing product or portrait photography for your feed or shop, Face Retouch gives you more control and finesse.Why we didn't add a manual brush option
We considered it. Seriously. A brush tool for blemish removal would give users choice, right? But choice is expensive in two ways: it complicates the interface, and it puts burden on the user. We'd have to add brush size, brush hardness, brush opacity controls. The learning curve climbs. Casual users bounce. And we'd solved the hard part already, the part that was actually limiting people from editing their photos. Why add complexity for the 5% of users who want surgical precision when 95% just want their skin to look decent? That decision shapes how Cleanr works across the board. We've picked intelligent defaults and built features that do what most people actually need. Auto Blemish, background removal with 15 presets instead of blank canvas, sky replacement with six curated options. Opinions built into the design, not decision trees.What we learned by shipping Auto Blemish to real people
Once Auto Blemish went live, the feedback split into two buckets. Faith creators and Christian social media users loved it. They could edit selfies and group photos for their communities without feeling like they were being vain or artificial. One user told us: 'I can post a photo I actually feel good about without spending an hour in Photoshop. That's freedom.' The second bucket was everyday users on Instagram, TikTok, and Shopify. Small business owners using Cleanr to clean up product photos. They appreciated not having to download Photoshop or learn Lightroom just to remove a blemish or two. What surprised us was how many people told us they'd downloaded other editors but bounced because they were watermarked, or demanded subscriptions with confusing credit systems. We'd shipped Cleanr free with Auto Blemish included, unlimited on the Free tier. No watermark. No nonsense. That positioning mattered more than the feature itself.Auto Blemish in the context of your edit workflow
Most people edit in sequence. They open a photo, maybe auto-enhance it (one tap for colour, exposure, sharpness), then handle specifics. For portraits, Auto Blemish lives early in that flow, right after enhancement and before any background work. If you're editing a selfie, you'll often run Auto Blemish, then use portrait blur (which also uses Vision face detection to isolate your subject) to soften the background and push focus to your face. If you're a small business owner shooting product photos with people in them, Auto Blemish gives you a professional baseline without requiring technical skill. On AI Pro, you can layer Face Retouch on top for more selective control, use HSL colour adjustments to fine-tune skin tone, or deploy selective adjust with a brush to brighten just the eyes or reshape a cheekbone. But the foundation is Auto Blemish, and it works because it doesn't ask you for anything except permission to improve the photo.The real question isn't whether Auto Blemish removes blemishes well. It's whether you've ever closed a photo editor halfway through because it was asking you to learn too much just to look your best. That's what we were trying to fix.