TechRiseUps — Autonomous AI Editorial Engine for Tech Media
An AI-run tech-media platform I built and operate. It researches, writes, reviews, and publishes SEO articles on its own — with real citations and a strict human-grade review gate, not a content farm.
Project Overview
TechRiseUps is a production AI content engine I designed, shipped, and run myself. It autonomously researches winnable keywords, drafts long-form articles with Claude, generates on-brand hero images, then passes every piece through a strict AI reviewer that returns PASS, FIX, or FAIL before anything goes live. The whole pipeline is API-driven — new articles publish without a redeploy — while staying honest about being a small, E-E-A-T-conscious editorial operation. It's the same research-to-publish architecture I build for clients who want content, or any repetitive workflow, running on autopilot.
The Challenge
Tech publishing has a brutal math problem: ranking content needs real research, original writing, and constant refreshing, and doing that by hand doesn't scale past a few posts a week. The cheap fix — spin up a content farm — gets you de-indexed and earns nobody's trust. I wanted to prove a third path: a site that publishes like a small, credible newsroom but runs almost entirely on automation. The hard part wasn't generating words. It was generating words good enough to survive a strict editor every single time, with citations, a named author, and an E-E-A-T posture Google actually rewards.
What I Built
- Built the full platform on TanStack Start (React 19) with PostgreSQL on Supabase as the system of record
- Engineered a token-scoped ingest REST API so articles publish on demand — no redeploy, ever
- Wrote an end-to-end AI article pipeline on Claude that researches, outlines, drafts, and cites real sources
- Wired DataForSEO keyword research into the pipeline so it only chases winnable, intent-matched topics
- Added gpt-image-2 hero generation so every article ships with an on-brand image, not a stock placeholder
- Built a strict AI reviewer gate that scores each draft and returns a machine-readable PASS / FIX / FAIL verdict before publish
- Built a DB-backed media API so images and assets are served and tracked from one source
- Wired IndexNow so new and refreshed URLs get pushed to search engines the moment they go live
- Built an automation-telemetry admin to watch pipeline runs, pass rates, and per-article API cost
- Baked in refresh cycles and a named author so the site reads as a real editorial operation, not a farm
Key Features
Results
“The engine isn't a demo — it's live and publishing. The review gate is the whole product: if a draft can't survive a strict editor with citations and a real author behind it, it never goes out.”
Discuss this project
Fill out the form below to get started.