Build yours.

A step-by-step guide to building your own identity store.

You do not need to be a developer. You need a computer, some patience, and a willingness to look at your own data honestly.

What this is

A personal recommendation engine. You feed it data about your life -- what you buy, listen to, read, watch, run, research -- and it builds a model of your taste. Then it recommends products that match that model.

Every ad platform does this already, but for advertisers. This does it for you. Same technology, opposite beneficiary. It runs as a website. You can share it, keep it private, or just use it as a personal reference for what you own and what you might want next.

What you will need

required

A computer

Mac, Windows, or Linux. You will be running terminal commands.

required

Node.js

Version 18 or later. Download from nodejs.org.

required

A GitHub account

Free. You will fork the repo here.

required

A Vercel account

Free. This is where your site gets hosted.

recommended

Claude Code

Build conversationally instead of editing files manually.

helpful

2-3 hours

For the initial setup and data import.

How to set it up

01

Fork the repository

Go to the GitHub repo and click Fork. This creates your own copy.

https://github.com/dominicbuckland-del/domshop
02

Clone it to your computer

Open your terminal and run:

git clone https://github.com/YOUR-USERNAME/domshop.git
cd domshop
npm install
03

Start the dev server

npm run dev

Open http://localhost:3000 in your browser.

04

Replace the products with yours

Open data/products.ts. Each product has these fields:

{
  name: 'Your product name',
  oneLiner: 'One sentence about why',
  whyILikeIt: 'The longer story',
  price: 2999, // in cents
  link: 'https://where-to-buy.com',
  category: 'work', // work, play, health, mind, etc
  image: 'https://your-image-url.jpg',
  recommendedBecause: 'Why the algorithm surfaced this',
}

Photograph your stuff, upload images, paste the URLs.

05

Update the algorithm data

Open data/algorithm.ts. Replace the data sources with your own. Start with Level 1 manual imports -- you do not need live integrations to launch.

06

Deploy

npm install -g vercel
vercel deploy --prod

Free hosting, free SSL, instant. If you have a custom domain, connect it in the Vercel dashboard.

How to connect your data

The quality of your recommendations depends entirely on the quality and honesty of your data. Here is how to export from each platform.

Spotify

  1. 1.Go to spotify.com/account > Privacy settings > Download your data
  2. 2.Request "Extended streaming history" (not the basic one)
  3. 3.Spotify will email you the download (can take up to 30 days)
  4. 4.The algorithm uses this for: listening hours, genres, mood patterns, device context

Strava

  1. 1.Go to strava.com > Settings > My Account > Download or Delete Your Account
  2. 2.Click "Request Your Archive"
  3. 3.Contains activities.csv with every workout: type, distance, duration, gear used
  4. 4.The algorithm uses this for: fitness patterns, gear wear cycles, training volume changes

Claude conversations

  1. 1.Go to claude.ai > Settings > Export Data
  2. 2.Download the JSON export of all your conversations
  3. 3.Richest data source: captures reasoning, not just clicks
  4. 4.The algorithm learns what you research, compare, and reject

Instagram

  1. 1.Settings > Accounts Centre > Your information and permissions > Download your information
  2. 2.Select: Saved posts, Following, Liked posts. Choose JSON format.
  3. 3.Download contains every post you saved or liked
  4. 4.The algorithm reads visual preference and brand affinity patterns

Browser bookmarks

  1. 1.Chrome: chrome://bookmarks > Three dots > Export bookmarks
  2. 2.Safari: File > Export Bookmarks
  3. 3.The algorithm looks for: product pages visited multiple times, wishlists, comparison shopping patterns

Email purchase receipts

  1. 1.Search Gmail for "order confirmation" or "receipt"
  2. 2.Or use Google Takeout (takeout.google.com) to export all Gmail data
  3. 3.Key data: what you bought, when, how much, whether you reordered
  4. 4.The algorithm uses purchase frequency and reorder patterns for replacement predictions