NorthCinder
An open-source plug-in that lets AI assistants shop for you honestly — comparing prices across stores and always asking before it actually buys anything.
🔗 Visit NorthCinder
Description
As AI assistants start being trusted with tasks like online shopping, two problems come up immediately: will it recommend the product that's actually best for you, or the one a retailer paid to promote — and will it ever spend your money without asking first? NorthCinder is built specifically to answer both with "no bias, and never without approval." It's a Model Context Protocol (MCP) server that any compatible AI agent can use to compare products across Shopify, WooCommerce, eBay, Etsy, and Amazon stores.
It runs entirely locally with no cloud service and no user accounts — the repository's own maintainers explicitly note they don't operate a NorthCinder service themselves, so there's no hidden backend collecting your data. Every comparison includes transparent source attribution so you can see where a recommendation came from, and mandatory buyer approval is built into the architecture, not just a suggestion — no purchase happens without it. It's built with TypeScript and Node.js in a pnpm workspace, includes documented ranking and trust specifications for full transparency, and has crossed 1,200 GitHub stars.
💬 Our review
The short version: for anyone letting an AI agent near their shopping, NorthCinder's local-first, no-cloud, mandatory-approval design is exactly the right set of guardrails, and it's rare to see a shopping-comparison tool built this deliberately privacy-first.
There isn't a direct competitor doing quite the same thing in the open-source MCP ecosystem yet — most "AI shopping" tools are commercial and cloud-hosted, which is the opposite trade-off: convenience and broader retailer coverage in exchange for your data flowing through someone else's servers. NorthCinder's limitation is coverage — Shopify, WooCommerce, eBay, Etsy, and Amazon is a solid start but won't catch every retailer, and as a young, single-repository project it depends on continued maintenance to keep pace with each platform's changes. For developers who want to add trustworthy, bias-free shopping comparison to their own AI agent setup without sending data anywhere, it's a strong open-source building block; anyone wanting a polished consumer-facing shopping assistant today should look at commercial alternatives instead.
💰 Pricing
📊 Global score
🤖 AI-enriched data
100% gratuit, licence MIT, exécution 100% locale sans compte ni service cloud.
Pros
100% local, aucun compte ni service cloud — confidentialité maximale
Approbation de l'acheteur obligatoire avant tout achat, intégrée à l'architecture
Attribution de source transparente pour chaque comparaison
Couvre Shopify, WooCommerce, eBay, Etsy et Amazon
Spécifications de classement et de confiance documentées publiquement
Cons
Couverture limitée à 5 plateformes, pas tous les retailers
Projet jeune, mono-mainteneur — dépendance à la continuité de maintenance
Pas d'alternative directe connue pour comparer objectivement le positionnement
Nécessite un agent compatible MCP pour être utilisé, pas un produit grand public autonome