A restaurant owner knows exactly what they sold last night. Very few know what they earned on it.

What they built

Recepta calls itself the copilot of the food trades, meaning restaurants, caterers, bakers and pâtissiers. It connects the things that decide whether a kitchen makes money, which are usually kept in different places or nowhere at all: supplier invoices, ingredient prices, recipes and their technical sheets, the cost of materials and production, and the margin on every dish.

The mechanism is simple. You photograph or upload a supplier invoice, Recepta reads it, and the new prices flow through into every recipe that uses those ingredients. The margin on each dish updates by itself. What comes out is a decision rather than a report: raise this price, change this recipe, switch this supplier.

The company says it is already in use in 40 establishments with more than a hundred users, spread across France from Lille, Arras, Amiens and Dunkirk to Paris, Lyon and Vittel, and that no customer has left since it launched. Most are independent restaurants, with caterers and bakeries making up the rest. It costs 75 euros a month for a single user. Its président is Gabriel Roussel, and the team is four people. The founder is twenty-three, trained professionally as a pâtissier, and worked in about a dozen restaurants before starting the company, where he found the same margin problem in every one. Recepta works with regional partners including the Inov'Alim food cluster and the junior enterprise of the Paul Bocuse culinary school.

Restaurant software is a crowded market. Toast, Lightspeed and, in France, Zelty run the till. MarketMan and Apicbase handle inventory and purchasing for larger groups, meez and Galley manage recipes, and Inpulse forecasts demand. Recepta's own comparison is blunt. Spreadsheets are manual and rarely up to date. Tills are strong on sales and weak on real costs. Enterprise systems are far too heavy for a small kitchen. Very little is built for the owner of a single restaurant who is on the floor at eight o'clock and doing the invoices at midnight.

The point: the person who lived the problem can now build the fix

For most of the history of software, the people who best understood a trade could not build tools for it. A pâtissier who spotted the same margin problem in twelve kitchens would have needed a technical co-founder, a funding round and a year before anyone could use anything. Most never got that far, so restaurant software was written by software people, usually for chains, and small kitchens were left with spreadsheets.

That constraint has mostly gone. With AI doing a large share of the building, a small team led by someone from the trade can ship a real product quickly, and the scarce ingredient is no longer engineering. It is knowing exactly which problem hurts, in which order, and what the owner will actually do about it at midnight.

Recepta's choice of entry point shows that knowledge. It does not start with the till, which everyone else fights over, or with a full back office nobody has time to set up. It starts with the invoice, because that is where margins quietly change. A supplier raises the price of butter, the cost of forty dishes moves, and the owner usually finds out months later, if ever. Catching that moment is worth more to a small kitchen than any dashboard.

The route to market reflects the same instinct. Rather than selling door to door to more than a million establishments, the company plans to go through the people restaurants already trust: accountants, till suppliers, food wholesalers, hotel schools and restaurant consultants.

Where else this works

This part is opinion rather than anything the company has said.

The formula is to look for trades where the owner works on the floor rather than at a desk, where existing software is either too heavy or blind to costs, and where someone who has done the job can now build the tool. Hair salons and their product costs. Garages and their parts margins. Florists and their perishable stock. Dental practices and their consumables. Small builders and their materials. Farms and their inputs. Print shops, bike shops, butchers. In each of them the person who understands the problem best is now, for the first time, able to fix it themselves.

Why it is bigger than it sounds

The short description is recipe costing software. The shift underneath it is who gets to build software at all.

  • The insider is now the founder. Someone from the trade no longer has to wait for a technical partner, which moves the advantage to whoever understands the work best.
  • The invoice is the right door. Margins move when supplier prices move, and catching that moment matters more to a small kitchen than anything the till can show.
  • Nobody built for the single site. Tills serve sales and enterprise tools serve chains, which leaves the independent owner as a large and underserved market.

What to watch

The measure is whether the churn stays at zero as the customer count moves from forty to four hundred. Early customers are often friends of the founder. Strangers who stay are proof.

The second is the accountants. If accounting firms start recommending Recepta to their restaurant clients, the company has found a way to reach a million kitchens without knocking on a million doors.

The third is how far the product reaches. Stock, inventory, hygiene and scheduling are the company's stated next steps, and eventually the till itself. Moving from margins into the till is where it meets the biggest competitors directly.