reddit · SideProject · 10-14 12:50 47 分 ← 返回列表

I built an app during my Erasmus because I was surrounded by art I didn’t understand. Co-founders left the project. Here’s where it is after a full rebuild

Why I started During my Erasmus I was surrounded by monuments and sculptures, went to tons of museums and travelled quite a lot. But I never really felt connected to what I was seeing, unless I paid for a tour guide every time. That’s where the idea for AnyArt came from: point your camera at a painting or monument and get its history, artist and symbolism. The first version I started by connecting third-party services and got a working MVP. But it was far from usable:

五维评分

付费意愿

需求真实度

个人可落地

竞争空间

新鲜度

AI 分析

痛点：游客看不懂眼前的画作与古迹，除非每次花钱请导游或买导览，缺乏低成本即时理解的替代方案。

目标用户：出国旅行的游客、博物馆参观者、艺术与历史爱好者。

证据：I never really felt connected to what I was seeing, unless I paid for a tour guide every time. / I had tried Meta Ads before and got exactly zero paying users, so this time I'm focusing on organic content.

变现：免费增值：免费每日识别次数，订阅解锁无限识别与深度解说；博物馆/旅游局B2B授权与导览内容打包；联盟分佣导览与门票。

MVP：直接调用视觉大模型API（GPT-4V/Gemini）做拍照识别，输出作品名、作者、年代、象征寓意，2周内做到识别<10秒并支持按作者/国家自动归类；先深耕单一城市或单一博物馆做SEO内容获客验证付费转化。

竞品：Google Lens、Google Arts & Culture、Smartify、ChatGPT/Gemini 视觉识别、博物馆官方语音导览App

摘要：独立开发者在Erasmus交换期间因看不懂艺术品而做出AnyArt：拍照识别画作与古迹，返回历史、作者与象征含义。重写后识别从近1分钟缩短到6-7秒，覆盖Google Maps上全部古迹及欧美主要博物馆藏品，期间两位联创离开。但作者自述此前Meta Ads投放获得“零付费用户”，现转为纯自然流量，付费模式尚未验证。