AI Tool
dailypriceguru.com
Live siteA price comparison platform where Claude extracts product specifications, normalizes them to 0-100 scores, and aggregates live pricing across multiple retailers.
Most comparison sites are affiliate funnels disguised as editorial. dailypriceguru.com runs a real AI pipeline — Claude extracts specs, normalizes them to 0–100 scores per category, and writes a Guru's Verdict that does the actual analytical work. Sub-second ISR loads keep the experience fast even while the pipeline runs async.
3
AI pipeline stages
<1s
Page load (ISR)
5+
Retailers tracked
Screenshots
The Problem
Product comparison sites are either affiliate-stuffed ad farms with fake "best of" rankings, or raw spec tables that require the user to do all the analytical work. There was a gap for something that does the actual comparison thinking — using AI to generate an honest verdict.
The Process
Built a Claude pipeline that takes a product name, fetches specs from retailer APIs, normalizes disparate spec formats into comparable 0-100 scores per category, then generates a "Guru's Verdict" narrative comparison. ISR on marketing pages keeps load times fast while the data pipeline runs asynchronously.
The Solution
A full-stack comparison platform with AI-powered spec extraction, multi-retailer price aggregation, normalized scoring, price history charts, and a "Guru's Verdict" AI comparison narrative for each product pair.
What Was Built
- Claude pipeline: spec extraction → normalized scoring → verdict
- Multi-retailer price aggregation with history charts
- ISR on public pages: sub-second loads
- Async AI processing with cached results
- Product referral monetization (no Stripe dependency)
Feature Breakdown
AI Core
- Automated spec extraction with 0–100 normalization scoring
- Guru Score: weighted overall product rating
- Guru's Verdict: AI-generated plain-English comparison narrative
- Explainable score breakdown (transparent why behind scores)
Price Intelligence
- Multi-retailer price aggregation (Amazon, Best Buy, Walmart+)
- 30/90/180-day price history charts per retailer
- 'Best Time to Buy' trend indicator
- Fake discount detector (price manipulation flagging)
Trust & Discovery
- Source citation with confidence badges
- No affiliate rankings — referral monetization only
- ISR on public pages: sub-second loads
- Async AI processing with cached results
Tech Stack
Source Code
Private repository
Behind the Build
The normalization pipeline was the trickiest engineering problem here — getting disparate spec formats (screen sizes in inches vs. cm, battery in mWh vs. mAh) into comparable 0-100 scores required a lot of category-specific logic. ISR was the right call for public pages; the async AI processing means the first load is fast even when the pipeline is still running.