HackerNews中文版

我最近推出了 SendStackr。最初的灵感来自于我在一家代理机构工作时的经历，这家机构管理着多个 Upwork 账户和一个“电话农场”业务（大约 50 台物理设备）。我们需要集中处理收到的电子邮件并转发短信（主要用于 2FA/OTP），但现有的工具要么贵得离谱，要么迫使我们为每个设备设置单独的 Webhook，要么在尝试通过标准的无代码封装器传输大量非结构化数据时超时。为了解决这个问题，我构建了自己的路由和提取引擎。我没有依赖标准的 API，而是完全用 Rust 编写了后端，构建了自定义的 SMTP 和 IMAP 服务器来原生接收电子邮件。它使用图 RAG 来读取非结构化 PDF、原始 HTML 电子邮件和 Webhook，提取所需的精确变量，并将它们作为严格映射的 JSON 返回。我还发布了一个 Android Node 应用程序来处理物理电话问题（短信和推送通知转发）。

查看原文

I recently launched SendStackr. The original itch came from my time working at an agency that managed multiple Upwork accounts and a "phone farming" operation (around 50 physical devices). We needed to centralize inbound emails and forward SMS (mostly for 2FA/OTPs), but the existing tools were either prohibitively expensive, forced us to set up individual webhooks per device, or timed out when trying to pipe heavy unstructured data through standard no-code wrappers.<p>To solve this, I built my own routing and extraction engine.<p>Instead of relying on standard APIs, I wrote the backend entirely in Rust, building custom SMTP and IMAP servers to ingest the emails natively. It uses Graph RAG to read unstructured PDFs, raw HTML emails, and webhooks, extracting the exact variables needed and returning them as strictly mapped JSON.<p>I also just released an Android Node app to handle the physical phone problem (SMS and push notification forwarding)

我用 Rust 构建了一个 AI 数据提取引擎和端到端加密的短信路由器。