Building an iPad PDF Reader with AI: What I Learned Beyond Generating Code

说明与用途

作者构建了一个 iPad PDF 阅读器，能够正确处理从右到左阅读、封面页和双页展开，提升了渲染质量，并建立了可重复的 AI 辅助开发流程。

问题

The author needed a PDF reader that correctly handled right‑to‑left reading, cover pages, and two‑page spreads on an iPad, but existing free viewers did not meet these requirements.

结果

The author successfully built a custom iPad PDF reader that met specific layout and reading‑direction requirements, improved rendering quality, and established a repeatable development and testing workflow using AI tools.

工作流

• Describe desired behavior to AI
• Generate or modify code with Gemma/ChatGPT
• Build the app in Xcode
• Run on Simulator and iPad
• Observe output and note issues
• Refine AI prompts with specific conditions and expected results
• Repeat build‑test‑observe cycle
• Add automated tests with shareable PDFs
• Prepare submission build with privacy policy and manifest
• Submit to App Store and monitor review status

证据

I read books I have scanned myself. I wanted a more comfortable way to read those PDFs on my iPad.

来源

My main frustration was how PDF readers handled right-to-left reading and cover pages.

来源

I began development in July 2026. My initial workflow involved two computers: On Windows, I used Cline with a local Gemma model running through LM Studio to generate and modify code. On a Mac, I built the app in Xcode and checked it in the Simulator and on a physical iPad.

来源

In an early version, the pages overlapped. I reported the problem and received a response saying it had been fixed. When I ran the app again, the pages were still misaligned, but in a different way.

来源

The app does that, but while reading, I found the changing text size distracting. I reconsidered the interface layout.

来源

I prepared shareable PDFs with known properties: Reading-direction and page-layout settings. Search terms with known occurrence counts. Large file sizes and high page counts.

来源

Review status checked on November 3, 2026 (Japan Standard Time): App Store Connect still showed version 1.0.4 (build 10) as “In Review.”

来源