📡 Idea Radar 每天 8:00 / 18:00 推送 · 18:00 生成精选
reddit · SideProject · 57 分 ← 返回列表

I moved to Germany, started counting my grocery receipts, and ended up building an app for it

I live in Essen and do most of my shopping at Lidl. I wanted to know one thing: where does the food money actually go? My bank says "Lidl 84 €". It can't tell me how much of that was cheese, or that the cheese got more expensive. The existing receipt apps I tried read the text but got the structure wrong: wrong year, columns shifted, three receipts glued into one. So I built Bonlo : photo of a receipt → every item in a category, checked against the total; a shopping list that builds

五维评分

付费意愿
52
需求真实度
45
个人可落地
88
竞争空间
45
新鲜度
58

AI 分析

痛点:银行账单只知道「Lidl 84€」,无法拆解到具体商品品类,也看不出某样东西涨价了;现有小票识别App结构读错(年份错、列错位、三张小票粘成一张)。

目标用户:想精细追踪食品开销与单品价格变化的欧洲(尤其德国)消费者,常去Lidl等折扣超市、对物价上涨敏感的家庭采购者。

证据:My bank says "Lidl 84 €". It can't tell me how much of that was cheese, or that the cheese got more expensive. / 10 free scans, then a small subscription, because every scan costs me real AI model money.

变现:订阅制:免费10次扫描做体验门槛,之后按月/年小额订阅(成本驱动型定价,覆盖AI视觉模型调用成本);后续可加品类价格趋势报告、家庭多设备同步等增值项。

MVP:2周MVP:SwiftUI/Flutter拍照上传小票 → 视觉大模型解析成结构化行项目JSON(品名/单价/数量/折扣)→ 按品类归类汇总并与历史价格对比 → 本地存SQLite不上传账号。先只做德语系超市(Lidl/Aldi/Rewe)的小票模板,用税率校验总额做兜底纠错。

竞品:Receipt Hog、Fetch Rewards、Expensify、各类个人记账App(Money Manager、MoneyWiz)

摘要:一位搬到德国Essen的独立开发者发现银行流水只显示「Lidl 84€」,无法拆到商品级别,且现有小票App识别结构经常出错,于是自己做了Bonlo:拍小票→按品类归类→与常购价对比,SwiftUI + Cloudflare Workers纯单人开发,无账号无广告。变现为10次免费扫描后的小额订阅,理由是每次扫描都真实消耗AI模型成本,但帖中未给出任何用户付费或收入验证,需求仍待证实。