DealForge autonomously sources, scores, and writes investment memos on venture deals. Stop manually hunting.
1,180+ deals tracked · 22 AI investment memos · Updated daily
Built a LinkedIn content intelligence tool for creators and founders, here's how it works under the hood
Solo founder here, been building SachaSoft for the past few months. It's a LinkedIn content intelligence platform that tracks \~200 top creators in real-time, classifies their posts with AI across 40+ industries, and surfaces winning formats before they saturate. Wanted to share the technical side because this sub appreciates builder posts. **The core insight that drove the product** Most LinkedIn tools are AI writers. You give them a topic, they spit out a generic post in someone else's voice. The output is interchangeable and the market is saturated. The actual problem creators and founders have isn't writing, it's knowing *what* to write. Format trends shift every few weeks per niche, and by the time the advice trickles down to YouTube tutorials, it's already dead. So instead of building another generator, I built a real-time intelligence layer: track top creators, measure post velocity at multiple intervals (10 min, 30 min, 1 hour, 3 hours), classify the breakout posts, and extract repeatable hook templates. The positioning is "intelligence not generation." That's the wedge. **The architecture** Stack is Node.js + React + MongoDB. The interesting part is the data pipeline: * Residential proxies geo-matched to each tracked creator's location * Fleet of headless Chrome instances with sticky sessions * Multiple LinkedIn accounts rotating through requests * Posts get re-scraped at increasing intervals to measure velocity decay against each creator's own baseline * Every post hits a classification pipeline tagging it by industry, format type, and hook structure * Hook templates extracted from breakout posts, clustered by niche, and surfaced in a weekly digest The infrastructure was the hardest part. LinkedIn aggressively blocks automated access, and getting to a stable state where accounts don't get flagged took months of iteration. That infrastructure layer is genuinely the moat, anyone can write a classifier on top of public data, but very few people can co
SachaSoft addresses a genuine pain point in the creator economy by pivoting from generic AI generation to data-driven content intelligence. While the technical execution of the scraping pipeline is impressive for a solo founder, the venture faces significant platform risk and lacks the traction data necessary for a higher valuation.