sainsburys-mcp
An MCP server that enables adding groceries to a Sainsbury's basket from a plain-text shopping list via natural language, using existing browser session for authentication.
README
sainsburys-cli
Build a Sainsbury's Groceries basket from a plain-text shopping list, from the command line.
It talks directly to Sainsbury's internal Groceries API and authenticates by borrowing the session from a browser you're already logged in to — so there's no password to store and no browser being driven around. It never places an order; it only fills the basket, ready for you to review and check out yourself.
⚠️ This uses an undocumented, unofficial API. It can break if Sainsbury's changes their site, and it's for personal use with your own account only.
Install
cd ~/code/sainsburys_cli
python3 -m venv .venv
.venv/bin/pip install -e .
This gives you a sainsburys command (at .venv/bin/sainsburys).
First-time setup: log in
-
Open your normal browser and log in at https://www.sainsburys.co.uk.
-
Leave that session logged in.
-
Check it works:
.venv/bin/sainsburys whoamiYou should see your name/email. The CLI reads the Sainsbury's cookies from your browser automatically. If you have several browsers, point it at one:
.venv/bin/sainsburys --browser firefox whoami
When your session later expires, just reload sainsburys.co.uk in the browser and
run any command with --refresh.
Build a basket from a list
Write a shopping list, one item per line:
# weekly shop
2 x semi skimmed milk
free range eggs x2
wholemeal bread
600g chicken breast
bananas
Quantities can go at the start (2 x milk, 3 bananas) or end (milk x2).
Blank lines and # comments are ignored.
Preview first (nothing is added):
.venv/bin/sainsburys add-list shopping.txt
You'll get a per-line report: ✓ confident matches, ? low-confidence or
out-of-stock (with alternatives shown), and ✗ no-match. When it looks right,
commit:
.venv/bin/sainsburys add-list shopping.txt --commit
Only confident, in-stock matches are added; anything flagged for review is left
for you to add manually (e.g. with search + add). Then open the basket on
the website to review and check out.
Other commands
.venv/bin/sainsburys search "oat milk" # find products + their ids
.venv/bin/sainsburys add 7977681 --qty 2 # add one product by id
.venv/bin/sainsburys basket # show the current basket (with item ids)
.venv/bin/sainsburys remove "oat milk" # remove an item by name (or [item ...] id)
remove matches on the product name; if more than one basket item matches
equally (e.g. remove "milk" with two milks) it lists them and does nothing, so
you can re-run with the specific [item ...] id or --all.
Add --json to any command for machine-readable output (handy when Claude is
driving it).
Using it with Claude Desktop
Claude Desktop drives this through a bundled MCP server (sainsburys-mcp) so
your partner can just chat — "add milk, eggs and bread to my Sainsbury's basket"
— with no terminal. See docs/CLAUDE_DESKTOP.md for
the one-time setup. (In Claude Code, use the CLI directly as below.)
Using it with Claude (CLI)
You don't have to pre-format anything. Paste a messy list into Claude — prose,
a copied spreadsheet, mixed units ("6 pints semi-skimmed, a dozen eggs, loaf of
wholemeal") — and ask it to build the basket. Claude interprets the list itself,
then uses search --json to find each product and add to put it in the
basket, stopping to ask you whenever an item is ambiguous (which size? which
brand? the usual one's out of stock — swap it?). It confirms the basket at the
end; you review and check out on the website. See CLAUDE.md for the exact
protocol Claude follows.
If reading browser cookies doesn't work
Some locked-down work machines (or Chrome on recent Windows) block reading the
cookie store. Fallback: copy the request cookies from your browser's dev tools
(Network tab → any groceries-api request → Cookie header, and the
wcauthtoken request header) and set them as env vars:
set -x SAINSBURYS_COOKIE "WC_AUTHENTICATION_123=...; JSESSIONID=..."
set -x SAINSBURYS_WCAUTHTOKEN "..."
Running the tests
.venv/bin/pip install pytest
.venv/bin/python -m pytest
The tests cover the list-parsing and product-matching logic (which run offline); the API calls themselves need a live logged-in session to exercise.
Recommended Servers
playwright-mcp
A Model Context Protocol server that enables LLMs to interact with web pages through structured accessibility snapshots without requiring vision models or screenshots.
Magic Component Platform (MCP)
An AI-powered tool that generates modern UI components from natural language descriptions, integrating with popular IDEs to streamline UI development workflow.
Audiense Insights MCP Server
Enables interaction with Audiense Insights accounts via the Model Context Protocol, facilitating the extraction and analysis of marketing insights and audience data including demographics, behavior, and influencer engagement.
VeyraX MCP
Single MCP tool to connect all your favorite tools: Gmail, Calendar and 40 more.
graphlit-mcp-server
The Model Context Protocol (MCP) Server enables integration between MCP clients and the Graphlit service. Ingest anything from Slack to Gmail to podcast feeds, in addition to web crawling, into a Graphlit project - and then retrieve relevant contents from the MCP client.
Kagi MCP Server
An MCP server that integrates Kagi search capabilities with Claude AI, enabling Claude to perform real-time web searches when answering questions that require up-to-date information.
E2B
Using MCP to run code via e2b.
Neon Database
MCP server for interacting with Neon Management API and databases
Exa Search
A Model Context Protocol (MCP) server lets AI assistants like Claude use the Exa AI Search API for web searches. This setup allows AI models to get real-time web information in a safe and controlled way.
Qdrant Server
This repository is an example of how to create a MCP server for Qdrant, a vector search engine.