A chat where you type as you speak â«how much did I spend on fish last quarter?»â and the aiGrocery advisor checks your receipts and answers with the exact amount.
Itâs not a chatbot that improvises generic advice: itâs an advisor who, before answering you, looks for the data in your own receipts.
Within aiGrocery, on Martaâs tab, you have an open chat. You type your question in natural language, and Marta starts with an already calculated summary of your last 12 months: how much youâve spent in total, the breakdown by 17 categories, spending per supermarket, the month-by-month total, your top 15 spending products, and your budget status in the current cycle (amount, spent, remaining, and projected close).
If what you ask isnât in that summary âa specific quarter, a single category, a store, or a productâ Marta doesnât make it up: she queries your data. She has two tools. `query_spend` filters your receipt lines by dates, category, store, or product and returns the total spent, number of lines, units, breakdown by category, and your top 10 spending products within that filter. `query_product` searches for a specific product and returns how many times youâve bought it, the total spent, the first and last purchase, and the unit price for the last 12 times, with the store for each. With that data in hand, she answers you in plain text, with the figure.
When a specific action suits you, Marta leaves it at the end of her response as a button: add a product to your shopping list, open that productâs detail page with its price history, or set your monthly budget. One tap and done, without leaving the chat.
The conversation is saved: Marta remembers the last exchanges, so you can ask «what about the previous month?» without repeating the context. And if you havenât scanned any receipts yet, she tells you and invites you to start there, because without receipts, there are no figures to query.
You ask as you speak
No filters or menus. «How much did I spend at Mercadona this year?», «Where is my money going?», «Compare this month with the last.» Marta understands the question and finds the data.
Your figures, not estimates
Answers come directly from the lines of your scanned receipts. The prompt explicitly forbids inventing numbers: if the data isn't available, it consults before answering.
On-demand queries, not just summaries
If the question goes beyond pre-calculated data, Marta launches a real query against your data, filtering by dates, category, store, or product, and responds with the result of that query.
Instant product history
Ask about coffee or milk and it tells you how many times you've bought it, how much you've spent, and the price per kilo or liter for the last 12 purchases, including the store for each.
Responds and suggests
At the end of the answer, action buttons appear: add to shopping list, open product details, or set your budget. Up to four per answer, with a single tap.
Remembers the conversation
The last questions and answers travel with each message, so follow-up questions work without needing to repeat the context.
Quick starts
Four ready-made questions above the text fieldâ'What do I spend most on?', 'How can I save?', 'Monthly summary', 'Compare with last month'âfor when you don't know where to start.
In your language
Marta responds in the language you use the app, with the tone of a practical advisor: brief, concrete, and with the number upfront.

Create your free account on aiGrocery, upload your latest supermarket receipt, and start asking. The more receipts, the more accurate the data.