Hi HN, I built a small open-source email assistant out of pure laziness. My inbox is full of noisy notifications, handover chains, and long threads that make it hard to find the real intent of an email. So I hacked together a system using Node-RED (for event-driven orchestration) and DeepSeek (for analysis). It turned into a tiny full-stack hack: Node-RED orchestrates the flows, DeepSeek does the analysis, and a lightweight frontend shows the results. New emails trigger a Node-RED flow automatically A two-layer LLM setup: first generates ultra-short summaries (one glance readable), then ensures valid JSON output Frontend parses the JSON to show key intent, info, and suggested reply Supports separate Chinese and English flows from the prompt level For my email volume, the AI API cost is minimal (roughly $0.2/day). It turned my inbox chaos into something manageable, Save me much time, and I thought others might find it useful too. npm: node-red-contrib-email-analysis-deepseek Node-RED Flow Library: https://flows.nodered.org/flow/ad230f600644964156ffb18db2f16... GitHub: https://github.com/witch-Judy/email-analysis-deepseek Would love feedback from the community. |