Show HN: I built a free minimalist eBook reader with Apple Intelligence powers (mskayyali.com)

Comma Reader is a minimalist, privacy-first eBook app for iOS that uses on-device Apple Intelligence to deliver smart reading features without sending data to the cloud. It offers automatic categorization and tagging, chapter summaries and explanations, highlighting, a private wall for quotes, and full offline functionality—so everything from NLP-based summaries to organization runs locally on your device. For the AI/ML community this is a good example of practical on-device NLP: classification, summarization and Q&A-style explanations implemented with Apple’s local inference stack (Apple Intelligence) instead of server-side LLMs. That design prioritizes user privacy and lowers data-exfiltration risks while trading off potential model capacity and update cadence compared with cloud models. It highlights a growing pattern of consumer apps shifting inference to edge devices, raising interesting engineering questions about model size, latency, energy use, and how to deliver capable NLP features under device constraints. Comma Reader demonstrates how privacy-preserving, offline-first AI can be integrated into everyday reading workflows without a server backend.