Guides, research notes and what we learn building models.
Research-backed list of AI features that drive retention and engagement in mobile apps. What users want, what they ignore, and how to prioritize AI features based on actual behavior data.
Read the article
Three paths to AI in Flutter. Cloud APIs via the http package, TensorFlow Lite for classical ML tasks, and on-device LLMs via llama.cpp for text generation. A practical comparison for Dart developers.
Read the article
Three paths to AI in your iOS app. CoreML for Apple's ecosystem, cloud APIs for capability, and on-device LLMs via llama.cpp for cost and privacy. A practical comparison for Swift developers.
Read the article
How to add AI features to React Native apps. Cloud API integration with fetch, on-device inference with llama.cpp bindings, and a practical migration path from one to the other.
Read the article
Your existing cloud AI API logs are a ready-made training dataset. How to extract, clean, and format API interaction logs into fine-tuning data for an on-device model.
Read the article
A practical comparison of the best small language models for mobile deployment. Llama 3.2, Gemma 3, Phi-4 Mini, and Qwen 2.5 evaluated for on-device inference via llama.cpp.
Read the articleA side-by-side comparison of Anthropic's Claude and OpenAI's GPT models for mobile app integration. Pricing, rate limits, capabilities, and when neither is the right answer.
Read the articlePrompt engineering is fast and flexible. Fine-tuning is accurate and cheap at scale. Here is the practical comparison for mobile developers deciding between the two approaches.
Read the articleRAG is the go-to solution for giving AI domain knowledge. But on mobile, RAG reintroduces the server dependency you are trying to eliminate. Fine-tuning bakes the knowledge into the model itself.
Read the articleGoogle's Gemini API offers aggressive pricing and native Android integration. Here's what the pricing actually looks like at scale, where the free tier ends, and when on-device models make more sense.
Read the articleHow to use Google's Gemma 3 models for on-device mobile AI. Model selection, fine-tuning with LoRA, GGUF export, and deployment via llama.cpp on iOS and Android.
Read the articleA complete guide to using Meta's Llama 3.2 1B and 3B models in mobile apps. Fine-tuning with LoRA, exporting to GGUF, and deploying on iOS and Android via llama.cpp.
Read the articleIntegrate llama.cpp into an Android app with Kotlin. Covers JNI bindings, Vulkan GPU acceleration, model loading, memory management and device testing.
Read the articleStep-by-step guide to integrating llama.cpp into an iOS app. Project setup, Metal GPU acceleration, model loading, token streaming, and memory management for production deployment.
Read the articleMeasured llama.cpp token speeds for 1B and 3B models on Snapdragon, Tensor, Exynos and Dimensity phones, plus the RAM tier each chip can carry.
Read the articleReal benchmark data for running LLMs on iPhones via llama.cpp. Token generation speeds, memory usage, and thermal behavior across iPhone models from the iPhone 12 to iPhone 16 Pro.
Read the articleA step-by-step migration plan for moving your mobile app from cloud AI APIs to on-device inference. Data extraction, fine-tuning, integration, testing, rollout, and monitoring.
Read the articleHow to build AI features that work without an internet connection. On-device models, offline-first architecture patterns, and the use cases where offline AI is not optional.
Read the articleArchitecture patterns for building a conversational AI assistant that runs entirely on the user's device. Model selection, conversation management, UI patterns, and production considerations.
Read the articleHow to choose the right model size for your mobile app. Capability breakdown, device requirements, quality benchmarks, and the fine-tuning factor that changes the math.
Read the articleHow to run language models directly on the user's phone in a React Native app. Setup, model loading, streaming generation, and cross-platform considerations using llama.rn.
Read the articleHow to build semantic search that runs entirely on the user's phone. Local embeddings, vector similarity, and natural language queries across user content without a server or API.
Read the articleThe complete unit economics breakdown for on-device AI vs cloud APIs. Fixed costs, variable costs, break-even analysis, and the financial model for scaling mobile AI features profitably.
Read the articleHow to build AI-powered drafting features that work without internet. Email replies, message suggestions, note expansion, and content templates generated entirely on the user's device.
Read the article