I spent almost 2 years building a tool that saves me 2 days a year — am I an idiot?

reddit SideProject · 原文链接 ↗ · 42 分

When I started, I couldn't write a line of Python. I'd spent 20 years in BI — SQL Server, SSIS, Qlik — and the idea of building a desktop app felt like stepping off a cliff. So I taught myself Python, automation testing, and Git from scratch. There were false starts — solutions that looked perfect on paper and collapsed in practice. My only stakeholder (me) was an utter nightmare — always asking for "one more feature" and "couldn't we just tweak this a little?" I decided it had to work on a Windows, Mac and Linux, even though I didn't own a Mac and rarely use Windows. I be

五维评分

AI 分析

痛点：把英国银行对账单 PDF 手工整理成可用于自评报税/记账的结构化数据，每年要耗掉约 2 天重复劳动

目标用户：英国个体经营者、自由职业者、小微企业与小型会计师事务所的记账/报税人员

证据：It reads UK bank statement PDFs and exports clean Excel, CSV, or JSON in seconds. / So it does save me 2 days a year when I'm doing my self-assessment tax return, but it's also taught me Python...

变现：开源核心 + 付费桌面/云端版（批量处理、多银行模板、自动分类与会计软件直连），或按份数/年订阅（如 £5-15/月）；也可向会计师事务所按席位授权

MVP：2 周内做一个只支持英国 Top 5 银行 PDF 的解析器：模板规则 + LLM 兜底处理未知版式，输出统一 CSV/Excel，附一个简单的交易分类规则表；先在自己报税流程里跑通再挂 Gumroad 卖 £19 一次性授权

竞品：Dext、Hubdoc、Docparser、Parseur、MoneyThumb、openstan

摘要：一位做了 20 年 BI 的开发者从零自学 Python，花了近 2 年做出开源离线工具 openstan，把英国银行对账单 PDF 解析成 Excel/CSV/JSON，只为在自己报税时省下每年约 2 天。帖子是纯粹的学习与发布分享，全文没有任何付费用户、收入或他人付费意愿的证据，需求目前只被作者本人验证。银行对账单 PDF 转结构化数据本身是会计自动化中的真实痛点，但该赛道已有 Dext、Hubdoc、Docparser 等成熟付费产品，独立开发者切入需要靠更垂直（单国别、离线隐私、低价一次性授权）来抢空间。