What kind of AI can be added to a mobile app?
- Image processing: defect detection, counting, product/license plate/document recognition, face and body landmarks via the camera
- Photo analysis: upload a photo to get results on damage, skin, plants, product quality and more
- LLM assistant: in-app chat, Q&A, text and quote generation
- RAG: an assistant that answers from your own documents and cites sources
- Voice: speech-to-text, voice commands
- Prediction and recommendations: sales, stock, personalized content recommendations
On-device or cloud?
On-device models (Core ML, Vision, MediaPipe, TensorFlow Lite) run fast without internet, and images never leave the phone; they are preferred for quality control and privacy-sensitive work. Large language models and heavy analysis run via a cloud API or a model on your own server. Most apps use both together.
Real projects
Inspectra is an on-device AI app that performs visual quality control on production lines using the iPhone camera. For Ege Auto Life, I developed a feature that analyzes vehicle damage from photos uploaded by customers. FaceMorph AI performs 3D face analysis and simulation on an Android tablet using a 478-point face mesh. Details: AI integration, FaceMorph AI.
Cost and privacy
With cloud LLMs, cost is incurred per request; it is reduced through caching, short prompts and choosing the right model. Personal data is anonymized and stored in compliance with KVKK (Turkey's data protection law); for sensitive work the model runs on the device or on your own server.
What I offer
Image processing
Detection and counting via camera.
Photo analysis
Upload, get results.
LLM assistant
In-app chat and text.
On-device AI
Offline, private, fast.
RAG
Answers from your own documents.
App Store
Publishing and review process.
Technologies I use
Example projects
A selection of live projects serving real users.
Inspectra — AI Quality Control & OEE
Edge-AI iOS app for camera-based visual quality inspection.
View
Ege Auto Life — Local SEO & AI Damage Analysis
Local SEO website and AI damage analysis for a paintless dent repair shop in Bornova, Izmir.
View
iPhone Security Scan — Spyware Detection
iOS app that scans for state-grade spyware, hidden tracking apps and jailbreak traces.
ViewHow we work
Free intro call
We clarify needs, existing systems and goals together.
Quote & roadmap
Scope, technology, timeline and budget within 24 hours.
Staged development
Weekly demos and a live test environment.
Launch & support
Go-live, training and ongoing maintenance.
Free pre-analysisThe first meeting and roadmap are completely free, with no obligation.
Staged delivery and approvalYou review each module on a test environment before we move on.
Written scope and rightsScope, delivery and source code rights are defined in the contract.
Frequently asked questions
How long does it take to build an AI-powered mobile app?
It depends on scope; first a prototype with a single AI feature is built and its accuracy measured, then the rest of the app is developed.
Can the model work without internet?
Yes. With Core ML, MediaPipe or TensorFlow Lite, the model runs on the device without internet.
Can a ChatGPT-like assistant be added?
Yes. An in-app chat assistant is added using a cloud LLM API or a model on your own server.
Have you built AI-powered apps before?
Yes. These projects include the on-device quality control app Inspectra, vehicle damage analysis from photos (Ege Auto Life) and the 3D face analysis app FaceMorph AI.
Is our data safe?
For sensitive work, images never leave the phone; when the cloud is used, data is anonymized and stored in compliance with KVKK.