Chapper ships iOS client for local LLMs

This is a strong “make local AI feel native on mobile” launch, and that matters because most LLM clients still treat iPhone as an afterthought. Chapper’s feature set is broad enough to be a real daily driver, not just a remote-control wrapper for desktop models.

• –The local-first positioning is the differentiator: it targets self-hosted LM Studio/Ollama users who want the couch-to-model workflow without handing data to a cloud service.
• –MCP support, URL context, and structured output push it beyond a simple chat app into a mobile AI operations surface.
• –The reasoning UI and `<think>` parsing are smart bets for users running newer reasoning models like Qwen3 and DeepSeek-R1.
• –Export options, personas, memory, and search make it feel closer to a serious knowledge tool than a toy messenger.
• –The risk is scope: a feature-rich client can get crowded fast, so polish and reliability will matter more than the checklist.