Analysis updated 2026-05-18
Ask an AI assistant to search for and add groceries to your yerevan-city.am cart by name.
Check your saved delivery addresses and cart total through a conversation instead of the website.
Learn how to build a Model Context Protocol server that connects an AI agent to a third-party site.
| tjvjk/yerevan-city-mcp | davidichalfyorov-wq/openxiv | kentjuno/kjaudiobook-v1 | |
|---|---|---|---|
| Stars | 33 | 33 | 33 |
| Language | TypeScript | TypeScript | TypeScript |
| Setup difficulty | moderate | hard | hard |
| Complexity | — | 5/5 | 4/5 |
| Audience | vibe coder | researcher | developer |
Figures from each repo's GitHub metadata at analysis time.
Requires building from source with npm and manually pointing your AI assistant's config at the built file.
yerevan-city-mcp lets you order groceries from the Armenian grocery site yerevan-city.am by just talking to an AI assistant such as Claude. Instead of opening the website and clicking through menus, you tell the assistant what you need, something like "order milk and bread," and it searches for the products and builds a shopping cart for you. The project is careful about one thing in particular: the assistant can search products, show your saved delivery addresses, and add or remove items from the cart, but it never pays or places the order itself. Checkout and payment always happen in your own browser or app, so you stay in control of the final step. Setup only requires confirming your phone number once through a text message code, and after that the assistant stays logged in for future requests without asking again. Under the hood this is a small server written in TypeScript that follows the Model Context Protocol, a standard way for AI assistants like Claude Code, Claude Desktop, Codex CLI, and a few other editor tools to connect to outside services. Installing it means downloading the project folder, running a build command, and then pointing your AI assistant's settings at the built file so it knows the server exists. Once connected, the assistant gains a handful of specific abilities: searching products by name, listing addresses, adding or removing cart items by piece or by weight, and checking the cart's contents and total price. The README is upfront that this is an educational, unofficial project built to show how an AI agent can talk to a third party website, and it is not connected to or approved by Yerevan City or City Supermarket in any way. It comes with no warranty. The license allows free personal and non-commercial use, but using it commercially requires reaching out to the author for a separate paid license. A short todo list at the end notes some unfinished edges, including how it behaves for phone numbers without an existing account and whether it works on Windows or Linux.
A tool that lets an AI assistant order groceries from yerevan-city.am for you, though you still confirm payment yourself.
Mainly TypeScript. The stack also includes TypeScript, Node.js, Model Context Protocol.
Setup difficulty is rated moderate, with roughly 30min to a first successful run.
Mainly vibe coder.
This repo across BitVibe Labs
Verify against the repo before relying on details.