The Mobile App Business with Fexingo: iOS, Android, and Consumer App Companies

In this episode, Lucas and Luna explore how mobile apps are bringing on-device AI to agriculture, specifically for real-time crop disease detection. They examine the case of Plantix, an app that uses computer vision models running locally on a smartphone to identify diseases in crops like rice, wheat, and tomatoes — without needing an internet connection. Lucas explains how this technology is particularly transformative for smallholder farmers in regions like India and sub-Saharan Africa, where connectivity is patchy and agricultural extension services are scarce. Luna questions the accuracy trade-offs of on-device versus cloud-based models, and they discuss how federated learning helps improve the model over time without compromising user privacy. The episode also touches on the broader implications: this isn't just a niche tool — it's a template for how on-device AI can democratize expertise in low-connectivity environments. The conversation includes a brief, organic mention of how listener support via buy me a coffee dot com slash fexingo helps keep the show ad-free and sustainable.

#Plantix #CropDiseaseDetection #OnDeviceAI #MobileApps #Agriculture #SmallholderFarmers #ComputerVision #FederatedLearning #India #SubSaharanAfrica #Business #Technology #FexingoBusiness #BusinessPodcast #Episode125 #MachineLearning #EdgeAI #FoodSecurity

Keep every episode free: buymeacoffee.com/fexingo

Episode 124 of The Mobile App Business with Fexingo explores a structural shift: mobile apps are on track to capture over 90% of consumer digital time by 2030, potentially making the mobile web browser an afterthought. Lucas and Luna break down the numbers from a recent Mary Meeker-style report, discuss what it means for developers building for iOS and Android, and examine why giants like Apple and Google have conflicting incentives to keep the browser alive. They also look at the counter-trend—progressive web apps and regulatory pressure—and ask whether the app store duopoly can sustain its grip without antitrust intervention. Specific data points include the current 87% app share of mobile minutes versus 13% for browsers, and the revenue implications for search-dependent companies like Google. A focused, data-driven conversation for anyone building or investing in the mobile ecosystem.

#MobileAppBusiness #iOS #Android #BrowserDecline #AppStoreDuopoly #DigitalTimeSpent #ProgressiveWebApps #Antitrust #Apple #Google #ConsumerApps #MobileFirst #WebVsApps #DeveloperEconomics #Safari #Chrome #BusinessAndTechnology #FexingoBusiness

Keep every episode free: buymeacoffee.com/fexingo

In this episode of The Mobile App Business, Lucas and Luna explore how mobile apps are using on-device AI to estimate the real-world dimensions of objects in real time. They dive into the technical mechanics behind apps like Apple's Measure and Google's ARCore-based tools, discussing how depth sensors, neural networks, and augmented reality overlays allow users to measure furniture, packages, and even body parts without a tape measure. The hosts examine a specific case: how a logistics startup used this technology to reduce return rates by 18% by letting customers measure boxes before shipping. They also touch on the challenges of accuracy in varying lighting conditions and the potential for integration with e-commerce platforms. Tune in for a focused look at a practical AI application that's quietly reshaping how we interact with the physical world through our phones.

#ObjectDimensionEstimation #OnDeviceAI #MobileApps #AugmentedReality #ARMeasure #AppleMeasure #GoogleARCore #RealTimeMeasurement #LogisticsTech #ECommerce #ComputerVision #DepthSensing #NeuralNetworks #FurnitureShopping #ShippingAccuracy #Business #Technology #FexingoBusiness

Keep every episode free: buymeacoffee.com/fexingo

Episode 122 of The Mobile App Business explores how mobile apps are leveraging on-device AI for real-time facial expression recognition. Lucas and Luna dive into the technology behind emotion detection, using a specific case: the meditation app Calm's integration of on-device expression recognition to personalize guided sessions. They discuss the machine learning models that run locally on iPhone and Android devices, privacy advantages over cloud-based processing, and challenges like accuracy across diverse demographics. The episode also touches on commercial applications in mental health, automotive safety, and retail, and examines how companies like Affectiva and Apple's ARKit are driving innovation. Listeners learn about the trade-offs between latency, privacy, and model size, and why this technology is becoming a standard feature in consumer apps by mid-2026.

#FacialExpressionRecognition #OnDeviceAI #MobileApps #EmotionDetection #MachineLearning #PrivacyTech #CalmApp #Affectiva #ARKit #MentalHealth #AutoSafety #RetailAnalytics #DeepLearning #iOS #Android #EdgeAI #FexingoBusiness #BusinessPodcast

Keep every episode free: buymeacoffee.com/fexingo

Lucas and Luna explore how mobile apps are using on-device AI to translate sign language into text and speech in real time. They focus on a specific case: Google's MediaPipe framework and its hand-landmark model, which tracks 21 points per hand 30 times per second entirely on the phone. The episode walks through how the landmark data feeds a gesture classifier, the challenge of distinguishing similar signs, and why on-device processing is critical for privacy and latency. Luna brings data on the global deaf community—over 70 million people—and Lucas explains why most sign language translation apps still struggle with grammar and regional dialects. They also touch on Apple's acquisition of a sign language recognition startup and the open-source Sign Language Datasets project. The conversation ends with Lucas wondering if full bidirectional translation—speech to sign and sign to speech—will ever feel natural, and Luna noting that the hardware is already there, it's the data that's lagging.

#OnDeviceAI #SignLanguage #RealTimeTranslation #GoogleMediaPipe #HandLandmarkModel #AccessibilityTech #MobileApps #DeafCommunity #GestureRecognition #MachineLearning #EdgeComputing #AppleAcquisition #OpenSource #SignLanguageDatasets #BusinessAndTechnology #FexingoBusiness #BusinessPodcast #LucasAndLuna

Keep every episode free: buymeacoffee.com/fexingo

On-device AI is transforming agriculture by enabling real-time pest detection straight from a farmer's phone. Lucas and Luna dive into how computer vision models trained on millions of labeled insect images can identify crop threats in seconds, even offline. They discuss a case study from a startup in Kenya that reduced pesticide use by 40 percent using a mobile app called FarmGuard, which runs entirely on-device to avoid latency and data costs. The episode covers the technical hurdles—training models for thousands of pest species across different crops and lighting conditions—and the broader implications for food security and smallholder farmers. Lucas explains how transfer learning and model quantization make it feasible to run complex vision models on mid-range Android phones, while Luna questions the accuracy trade-offs and data collection challenges. It's a grounded look at how pocket-sized AI is changing the economics of farming.

#OnDeviceAI #MobileApps #PestDetection #AgricultureTech #ComputerVision #FarmGuard #Kenya #SmallholderFarmers #FoodSecurity #TransferLearning #ModelQuantization #OfflineAI #DeepLearning #Sustainability #BusinessAndTechnology #FexingoBusiness #BusinessPodcast #TechPodcast

Keep every episode free: buymeacoffee.com/fexingo

Lucas and Luna explore how smartphone apps are using on-device AI to screen for skin cancer in real time. They focus on the specific case of the SkinCheck app, which uses a convolutional neural network trained on over 100,000 dermoscopic images to classify moles and lesions with reported sensitivity above 95 percent. The hosts discuss the technical trade-offs of on-device inference versus cloud-based analysis, the regulatory landscape including FDA clearance pathways, and how this technology is being deployed in dermatology deserts where access to specialists is limited. They also examine the limitations—false positives, dataset bias toward lighter skin tones, and the risk of self-diagnosis replacing professional care. A concrete, numbers-driven look at a high-stakes application of on-device AI.

#OnDeviceAI #SkinCancer #SkinCheckApp #Dermatology #HealthTech #MobileApps #AI #MachineLearning #CNN #FDA #Telemedicine #Diagnostics #HealthcareInnovation #DeepLearning #ImageRecognition #Business #FexingoBusiness #BusinessPodcast

Keep every episode free: buymeacoffee.com/fexingo

Episode 118 of The Mobile App Business dives into the growing niche of on-device AI for wildlife identification. Lucas and Luna break down how apps like iNaturalist and Seek by iNaturalist use lightweight convolutional neural networks to identify animals, birds, insects, and plants from a phone camera — all without a cloud connection. They discuss the technical trade-offs: why Apple's Core ML and Google's ML Kit enable offline inference, how training datasets like iNaturalist's 100 million+ observations are curated, and why latency matters when a bird might fly away in seconds. The hosts also explore the business model: free apps supported by partnerships with conservation groups, and premium tiers for detailed ecological data. Specific numbers: the iNaturalist dataset contains over 100 million observations of 400,000+ species, and offline identification accuracy exceeds 85% for common species. The episode contrasts this approach with cloud-dependent alternatives like Google Lens, highlighting where on-device AI wins — especially for users in remote areas without cellular service. A concrete example: the Merlin Bird ID app by Cornell Lab of Ornithology uses on-device AI to identify birds by sound and image, processing a 10-second recording in under two seconds on a recent iPhone. The hosts also touch on the competitive landscape, including startups like PictureThis and PlantSnap, and the rising importance of privacy for users who don't want their location data sent to cloud servers.

#MobileAppBusiness #OnDeviceAI #WildlifeIdentification #iNaturalist #MerlinBirdID #CoreML #MLKit #MachineLearning #ConservationTech #ComputerVision #OfflineAI #Privacy #BusinessAndTechnology #AppDevelopment #PictureThis #PlantSnap #FexingoBusiness #BusinessPodcast

Keep every episode free: buymeacoffee.com/fexingo

Episode 117 of The Mobile App Business with Fexingo explores how on-device AI is enabling real-time breath analysis on smartphones. Lucas and Luna examine the technology behind apps that measure respiratory rate, detect carbon dioxide levels, and even screen for conditions like asthma or COPD using just a phone's microphone. They discuss the AI models that filter out background noise, the accuracy challenges compared to medical devices, and the privacy advantages of processing everything on-device. The episode also covers the business landscape: startups like BreathResearch and ResApp Health, FDA clearance pathways, and how app makers navigate the line between wellness tracking and medical diagnostics. Listeners learn about the specific signal processing techniques, the datasets used to train these models, and why Apple and Samsung are investing in respiratory health features. A concrete example is the phone-based spirometry test that measures forced expiratory volume in one second (FEV1) using only the microphone and on-device machine learning.

#OnDeviceAI #BreathAnalysis #MobileHealth #RespiratoryHealth #AIinHealthcare #AsthmaTech #COPD #Spirometry #DigitalHealth #WearableAI #PrivacyFirst #HealthTechStartups #FDA #AppleHealth #SamsungHealth #MachineLearning #Business #BusinessPodcast

Keep every episode free: buymeacoffee.com/fexingo

In this episode of The Mobile App Business, Lucas and Luna explore how on-device AI is transforming parking spot detection. They break down the specific technical approach used by apps like SpotHero and ParkMobile, which rely on convolutional neural networks and smartphone cameras to identify vacant spots in real time, without sending video to the cloud. The discussion covers the training data challenge — how companies collect dashcam footage to teach models to recognize curbs, lines, and car shapes — and the privacy advantage of keeping all processing on the phone. Lucas walks through the typical pipeline: image capture, edge detection, object classification, and a confidence threshold that determines whether a spot is marked as free or occupied. Luna questions the edge cases: what happens with shadows, motorcycles, or snow? They compare this to cloud-based alternatives and explain why latency and user trust are driving the shift to on-device. The episode also touches on how this feeds into smart city initiatives and navigation app integrations.

#MobileAppBusiness #OnDeviceAI #ParkingSpotDetection #ComputerVision #EdgeAI #SpotHero #ParkMobile #CNN #RealTimeProcessing #Privacy #SmartParking #UrbanTech #iOS #Android #AppDevelopment #BusinessAndTechnology #FexingoBusiness #BusinessPodcast

Keep every episode free: buymeacoffee.com/fexingo