Talk to the menu. Order from the table.

Guests scan a QR, ask the AI about allergens, spice, or what's actually vegetarian, and order from the table. Every answer comes from verified menu data.

Get a demo View on GitHub

Nishys Neerdosa · Table 12

What goes well with neer dosa?

Most guests pair it with the chicken curry. Soft dosa, rich coconut gravy, straight from the kitchen. Want me to add both?

Order Neer Dosa + Chicken Curry · ₹240AddSee more pairings

Yes, add both

Added to your order. How about a goli soda or some buttermilk to go with it?

Goli soda it is

Ask about any dish…

One clean schema, everything on top

Verified menu data powers the chat, the ordering, the analytics, and the agents.

AI chat that guests trust

Allergen and dietary answers come from verified data, never guesses.

QR menus + table ordering

Guests order from the table; tickets land in the kitchen.

Demand analytics

What guests view, ask about, and can't find.

Agent-ready

An MCP server lets AI assistants browse menus and place orders.

Accuracy

Price accuracy and dietary-marker fidelity are verified per item, with confidence scoring and human review where the extractor is unsure.

How it works

From a photo of the menu to guests ordering at the table, in three steps.

menu-card.jpg

menu.json · 42 items

{
  "name": "Neer Dosa",
  "price": { "plain": 120, "with_curry": 240 },
  "diet": ["vegan", "gluten_free"],
  "allergens": [],
  "confidence": 0.97,
  "review": "verified"
}

For developers

Open-source, top to bottom

The extraction engine, schema, and serving stack are AGPL-3.0.

Read it, run it, self-host it

Every layer is in the open: how menus are extracted, how items are verified, and how they are served to guests and agents.

</>MenuHero on GitHubAGPL-3.0

Building a POS or ordering platform?

Talk to us about embedding MenuHero: one verified schema behind your menus, ordering, and agents.

Get a demo

Frequently asked questions

Answers to common questions about MenuHero. Anything else, write to hello@menuhero.ai.

Where do the AI's answers come from?

From your verified menu data, never guesses. Allergen and dietary answers only say what the data supports.

How do I get my menu into MenuHero?

Send a photo of the laminated card, a PDF, or a POS export. We structure every item, price variant, and dietary marker.

How accurate is the structured data?

Price accuracy and dietary-marker fidelity are verified per item, with confidence scoring and human review where the extractor is unsure.

Is MenuHero open source?

Yes. The extraction engine, schema, and serving stack are AGPL-3.0. The code is on GitHub.

Can AI assistants use it?

Yes. An MCP server lets AI assistants browse menus and place orders.

Can I embed MenuHero in my POS or ordering platform?

Talk to us. Book a demo and we will walk through what embedding looks like for your stack.