这个网站是怎么做出来的
对话提需求
写代码
GitHub
自动部署
上线
文档
手写代码
跑测试
Review
自动部署
全程不用手写一行代码，我只负责提需求和验收；每次改动的细节都记录在下面的更新日志里。源代码在 GitHub →
网站的技术架构
服务端是一个 Flask + SQLite 的单体应用，没有独立前端构建，页面都是服务端用 Jinja2 直接渲染出来的； 用 gunicorn（1 进程 4 线程）跑在 Railway 新加坡区域的一个 Docker 容器里。图片用 Pillow 处理（生成分享图、 压缩上传图片），视频用 ffmpeg 压缩并抽封面帧。数据库、用户照片、小说封面/立绘/视频、豆瓣封面缓存都存在 Railway 的持久化 Volume 里，图片和静态资源通过边缘 CDN 缓存；豆瓣书影信息靠抓取网页拿到，AI 截图识别接的 是 Anthropic 的 Claude API。
Changelog
中文A record of this site's development from scratch to now (screenshots are freshly regenerated to reflect the current UI, not the original at the time of each change).
Code volume: entries from before this session (marked "estimated") are rough figures reconstructed in hindsight; every entry from this one onward is measured precisely at the time of the change.
2026-08-04
1 updates · ~15 lines that day-
Chapter editor now matches the reader view's width and font size
+15 linesThe chapter content textarea was using the generic form style, capped at 560px wide -- much narrower than the actual reading page, making it hard to judge real layout while writing. Now the editor and the reader view line up on width, font size, and line height (both 870px wide, 16px text, 1.9x line height), so what you see while editing is close to what readers will see. Hit a CSS specificity snag along the way: the new font-size rule was silently overridden by an earlier, more specific generic rule (`.form textarea`) until switched to a more specific selector.