The poster used Claude on Mac and iOS to plan and order a week of groceries, with partial success, and will keep using it up to the grocery-list step.
The poster used Claude on Mac and iOS to plan and order a week of groceries, with partial success, and will keep using it up to the grocery-list step.
The poster used Claude on iOS to turn their weight, height, activity level, and goal weight into macro targets, then had it ask follow-up questions, build a week of meals, and give macros and portion sizes for each one. They turned that into an Excel shopping list, moved it to Claude on Mac because iOS lacked full MCP support, and then tried to have Claude shop on H-E-B using recent-order history and live website access, stepping in when it stalled or chose badly.
Meal planning and list-making worked, but cart-building was only partly successful: Claude could navigate H-E-B and add some items, yet often picked wrong products, got stuck on pages, and needed repeated prompting. The poster said they would keep using it through the grocery-list stage, not full ordering.
The main catch was platform support: they needed the Mac app for fuller MCP/browser control, while the iOS app was enough for planning and list generation.
Plan dinners for [number] days for [number] people with these preferences: [diet, budget, dislikes]. Then build the combined grocery list, and add the items to my cart at [grocery store website]. Stop before checkout and show me the cart, substitutions, and total for approval.
Score every demo request against your ICP, update HubSpot and ping the right rep in Slack within a minute of the form submit.
Every new Zendesk ticket gets a category, priority and sentiment tag plus a draft reply in an internal note before an agent opens it.
Copilot and Claude review each PR for bugs, missing tests and risky changes before a human looks, so reviewers can focus on design.