MenuSnap

How it works

1. Take a photo of any restaurant menu, or upload one from your phone.
2. Each dish gets outlined where it appears on the photo as the AI reads the menu.
3. Tap any outline to see an AI-generated photo of what that dish probably looks like — anchored where you tapped, so the menu stays as your reference.
4. Tap "Share menu" to send the snapped menu to a friend. They open it on their phone and see the same dishes ready to tap.

When to use it

• You're traveling and the menu is in a language you don't read.
• You see a dish name you've never heard of (sweetbread? pâté en croûte? cavatappi?) and want to know what to expect.
• You're ordering takeout from an unfamiliar restaurant and want to picture the dishes before you commit.
• You want to share a menu with a friend the same way you'd share a Spotify song.

Frequently asked

Is it really free?

Yes. No signup, no account, no credit card. There's a tip jar if you like it. The service runs on free tiers, which means there's a daily ceiling on how many menus the world can read; if we hit that, you'll see a clear "back at midnight UTC" message instead of a broken-looking error.

Are the dish images real?

No — they're AI-generated based on the menu description. They're a visual guess, not what you'll actually be served. Use them to picture the dish, not to predict the plating.

Does it work on my phone?

Yes. It's built phone-first. On iPhone, open the URL in Safari, tap Share → Add to Home Screen, and it installs as an app icon. Same on Android Chrome.

Does it handle non-English menus?

Yes. The OCR layer reads dozens of languages. You take a photo, dishes get outlined wherever they are, you tap. The dish images are still generated from the original description, so they'll be roughly accurate to whatever the menu says.

Will the AI image match what I'm served?

Sometimes very closely; sometimes hilariously not. Restaurants vary a lot in plating and ingredients. Treat it like a thumbnail, not a contract.

About

MenuSnap is an indie project inspired by Andrej Karpathy's MenuGen. It uses Google Gemini 2.5 Flash for OCR and bounding-box detection, and Cloudflare Workers AI Flux Schnell for dish image generation. Free, no signup, runs as a Progressive Web App on iOS and Android.