StartupWire.in
An autonomous curation engine compiling startup updates, VC news, and tech announcements using Gemini filters and pgvector deduplication.
Filter Niche Clutter
To provide developers and tech founders with concise, clean summaries of startup launches and artificial intelligence developments, stripped of marketing fluff and clickbait.
Autonomous Curation
A self-sustaining content asset that executes scraping, curation, validation, indexing, and newsletter publication with zero daily human administration needed.
Programmatic SEO
Leveraging structured schemas, speed optimizations, and long-tail query indexings to build organic traffic pathways for startup keywords.
Core Crawler & AI pipeline
Understanding the pipeline that ingests, summarizes, deduplicates, and commits data.
1. RSS Scraper Daemon
Every 4 hours, a scheduled Node worker triggers. It traverses 40+ verified RSS endpoints (TechCrunch, VentureBeat, Hacker News, Product Hunt), parsing XML schemas and checking URL status codes.
2. Gemini Summarization Agent
Candidate feeds are processed using Google Gemini 1.5 Flash. The agent reads the article body, discards advertisements, writes concise summaries, assigns category taxonomies, and extracts key entities.
3. pgvector Cosine similarity filter
To prevent duplicate entries reporting the same PR, article titles are converted into 768-dimension vectors. Supabase calculates the cosine distance between the candidate and all YTD posts, rejecting matching items.
[RSS Feed Feeder Nodes]
|
v (triggered every 4h)
[Node.js Scraper Daemon]
|
v (Zod Schema Sanitizer)
[Prompt Injection Payload]
|
v
[Google Gemini 1.5 Flash API]
|
v (Structured JSON output)
[pgvector Embedding Lookup]
|
+--> Distance < 0.15 ? [REJECT DUPLICATE]
|
v (Distance >= 0.15)
[Supabase PostgreSQL DB]
|
v (Webhook Trigger)
[Edge Cache Sitemap Rebuild]Performance & SEO Engine
How StartupWire achieves under 200ms page load latencies while crawling organic rankings.
Serving pre-rendered static semantic HTML content via Vercel Edge networks. Edge endpoints revalidate caches instantly using Supabase database triggers.
Auto-injecting detailed NewsArticle, FAQ, and Breadcrumb structured schemas on every news path to enable rich-snippets indexing on Search engines.
Using aspect-ratio bounds and next-gen compression rules to keep Cumulative Layout Shift (CLS) at 0, passing Core Web Vitals checks.
Future Growth & Monetization
Roadmap targets, API integrations, and developer sponsorship channels.
- Verify pgvector threshold logic in staging (Completed)
- Deploy static site Edge cache revalidation patterns (Completed)
- Build dynamic newsletter digest worker pipeline (In Progress)
- Expose API endpoints for aggregated startup news feeds (Backlog)
StartupWire rejects generic ad networks like Google AdSense to preserve loading speeds and user attention. Instead, the roadmap targets micro-sponsorship configurations (text ads for developer products) and subscriptions for premium aggregated search indexing APIs.