Brett ZienowiczSay hello
← home/case study

RoboRatings

A data-driven shopping platform for robot vacuums, built solo and end-to-end with AI: a structured product database, programmatic SEO, and automated social distribution.

2025 - Now · Atlanta
Next.jsProgrammatic SEOAI contentAutomated socialRemotion video
Claude Code
Built with
100+
Products analyzed
5K+
Data points tracked
8+
Brands covered
01

A shopping platform, built solo with AI.

RoboRatings is a robot-vacuum shopping platform. It helps shoppers find the right vacuum through data-driven rankings, side-by-side comparisons, a personalized quiz, educational content, and live pricing across retailers.

Like most review sites, it earns affiliate commissions from retailers like Amazon - but it's built with the rigor and approach of a software product, not a thin content play. The website, the product database, the content engine, the live price ingestion, and the social distribution were all built solo using Claude Code. Since the start of 2026 it has grown from essentially no traffic to thousands of visitors per month through SEO and GEO.

RoboRatings homepage
Build
Solo + AI
Built with Claude Code by one operator
Catalog
100+
Products scored on real specs
Traction
Live + earning
Affiliate revenue, growing organically
02

Claude Code wrote most of the platform.

problem
The top robot vacuum review sites are managed by a team of people. Competing solo means finding leverage somewhere else.
approach
Build data-driven tools that genuinely help people shop, and use Claude to build them. Claude Code handles the engineering so I can focus on what to build and why it helps the shopper.
outcome
The result is a platform built and run by one person, with AI taking on work that would typically be spread across a small team. New features ship same-day, and content scales without a content team.
03

A programmatic SEO engine grounded in real product data.

A product knowledge base (specs, features, pricing history, review extracts, category context) feeds a set of page templates: individual product reviews, head-to-head comparisons, category guides, best-for lists, and informational articles.

Every page is generated from the same grounded data, so ratings and specs stay consistent across the site. Rankings come from a structured scoring model run against real specs and review data, not arbitrary picks. The structured data approach allows content to scale automatically as new products and brands are added to the database.

RoboRatings programmatic SEO page templates
Structured
5K+ data points
One product database powering every page
Page types
5+ formats
Generated from the same data
Live data
Auto-refresh
Prices and availability stay current
04

Built to rank on Google and get cited by AI.

Ranking well is its own discipline: page quality, schema, internal linking, search intent, and increasingly how AI search engines read and cite a site. Staying on top of all of it is constant work.

I work the site regularly with Claude and a set of dedicated SEO skills: auditing pages, surfacing opportunities, and implementing fixes and new features directly in the codebase. The same approach drives GEO, structuring content so AI search engines like ChatGPT and Perplexity can surface and cite it. Because the tool that finds the issue also writes the fix, improvements ship quickly.

RoboRatings SEO and GEO optimization
05

A self-hosted scheduler that keeps the accounts posting.

Social is a funnel back to the site, but posting across accounts by hand eats hours. I run a self-hosted Postiz scheduler on Railway that posts two kinds of content: educational/informational posts written by Claude, and real-time deal posts that fire when a vacuum tracked on the site drops in price.

The deal posts pull live pricing from the site and render a programmatic Open Graph image tied to each URL, so the post looks designed without a manual design step. Caption copy is generated to match each platform's format.

RoboRatings Postiz social scheduler
06

Turning a product spec sheet into a branded video.

Video drives a disproportionate share of product discovery online, but creating video for every product is expensive. I built an early MVP of a Remotion pipeline that takes a product's knowledge base and manufacturer video clips, generates a script and narration, generates custom scenes, and renders a branded video ready for YouTube and elsewhere.

Because it's built in Remotion, every render shares the same motion-design language, so the brand stays consistent even as the process gets automated. It's the newest and least-finished piece of the system, and a clear area to push next.

Next case study

Matterform

read case study →

A mid six-figure private-label home goods brand run solo across five channels, with in-house manufacturing and a custom ERP.

Let's build something cool.

Open to new opportunities in AI, Product + eCommerce. Atlanta-based, remote only. Best at 0→1 product work with AI doing real work under the hood.

or drop a note →