An illustrated menu, even without a photo shoot

A dish name in a foreign language tells the guest nothing. iAlacarte generates one image per dish from its name, recipe and type, checks it, and displays it with a clear notice: it is an illustration. As soon as you have a real photo, it takes its place.

Preview

What the guest sees

On each dish page, the image is accompanied by the notice displayed by the app, in the guest's language. Here is a real dish from our demo menu:

Linguine alle vongole

Indicative image — generated for illustration. The actual dish served may differ.

Generation

How the image is produced

From the dish details: the dish name, recipe, type (pizza, gratin, salad, cocktail…) and section describe the format served: a whole baked camembert, a parmigiana in its gratin dish, a pastis with its water carafe.
Checked: a vision model compares the resulting image with the dish name and recipe; an image that does not match is regenerated, up to a set number of attempts.
Stored: an image produced for a given description is reused as is, across all menus and each time yours is analysed again.
Branded drinks: common bottles (soft drinks, water, beer) have a shared image in a catalogue, the same across all menus.
On first view: a dish image is generated the first time a guest opens its page, then served instantly.

Your photos

Your own photos, dish by dish

From the dish editor, you can upload your photo, regenerate the AI image or delete it. A photo you upload is always shown instead of the generated image, and it is kept if the menu is analysed again.

For an entire venue, three modes are available: AI images, uploaded photos only, or no images.

Limitations

An illustration, not a commitment

The image is indicative and the guest is informed on every dish page; the dish served may differ.
A dish we know nothing about (an add-on with no known ingredient, for example) has no image: no image is better than an invention.
Generation is probabilistic: two dishes with the same description share the same image, but two different descriptions may produce different results.
The restaurant owner remains responsible for consistency between the published visual and the dish served; their own photos always take priority.

FAQ

Frequently asked questions about AI photos

Are dish photos real or AI-generated?

Both exist. Without a photo from you, the image is generated by AI from the dish details and marked as indicative. A photo you upload replaces the generated image.

Is the guest told that the image is generated?

Yes. Each dish page displays the notice “Indicative image — generated to illustrate the dish. The dish served may differ.” in their language.

How is the image checked?

A vision model compares the image with the dish name and recipe before it is served. An image that does not match is regenerated, up to a set number of attempts; an unknown dish has no image.

Can I mix my photos and AI images?

Yes, dish by dish: your photos for signature dishes, AI images for the rest. Your photos are kept with each new menu analysis.

How long does it take to illustrate the whole menu?

Each image is generated in a few seconds, the first time the dish page is opened, then stored. There is nothing for you to start or wait for.

Illustrate my menu

Upload your menu: each dish gets its image, checked and marked as indicative, which your photos can replace whenever you want.

Create my account View the iAlacarte landing page

Reminder — AI-generated images are indicative. They do not constitute a commitment to presentation. The restaurant owner remains responsible for consistency between the published visual and the dish served.