Research · 16 August 2026

This is the menu. This is the menu a coeliac is allowed to order from.

We read every dish on 4,857 restaurant menus across Ireland and the UK and tagged each one against the restaurant’s own allergen information. It lets us answer a question nobody has had the data for: not whether restaurants cater for coeliacs, but how much of the menu is left once they do.

What that looks like

Café en Seine in Dublin lists 20 main courses. 3 of them are gluten-free, and one of those is a bowl of olives.

Nothing here is a criticism of Café en Seine — they publish their allergen information, which already puts them ahead of most. That is exactly why they can be used as the example. Their page shows the same thing dish by dish.

And the UK is worse than Ireland — at saying so, not at cooking

The paperwork gap is 17 points. The food gap is 5. So among restaurants that publish enough to be judged, the two countries look much alike — almost the whole of the difference a coeliac actually experiences is a publishing difference, not a cooking one.

“Gluten-free” often means chips

Of the restaurants offering anything gluten-free, this is the share whose only gluten-free items are sides, sauces or puddings — no main course at all. Someone reading “we have gluten-free options” has between a third and a half chance of that meaning a bowl of chips.

City by city

The share of all live restaurants where a diner can confirm at least one gluten-free main course before leaving the house.

How this was measured

• It measures what restaurants publish, not what kitchens can do. A restaurant with no allergen sheet may cater superbly for coeliacs — we simply cannot tell, and neither can a diner deciding where to go. That findability is the thing being counted here.
• Every tag comes from the restaurant. We read the menu and allergen information each business publishes on its own website. We never infer that a dish is free from one of the allergens that can kill — those are only ever counted when the restaurant says so itself.
• Waterford’s percentage is real, its ranking isn’t. Only 35 of its restaurants publish enough to judge, against more than a thousand in Dublin. Read it as a number, not a league position.
• There are no dairy-free figures here on purpose. Our dairy-free tagging is deliberately cautious — we only apply it where a restaurant published a dairy column of its own — which makes it a sound safety rule and a misleading statistic. We would rather publish less than publish something we would have to footnote.
• The one-in-seven figure is the best case, not the average. It counts only restaurants that publish allergen information AND have something gluten-free. Restaurants with nothing gluten-free are excluded, and so are the majority that publish nothing at all — at those, the honest answer is that nobody can tell.
• There are no figures here for nuts, egg or the other allergens that can kill. We never infer that a dish is free from those — they are counted only where a restaurant says so itself — so any share we published would be far below the truth. Same reason as the dairy-free figures above: right as a safety rule, wrong as a statistic.
• Figures measured on 16 August 2026 across 4,857 live restaurants in Dublin, Cork, Galway, Limerick, Waterford, Belfast and Edinburgh.