A guest in our showroom on Zeltweg 74 held up the menu from an Italian restaurant last week and asked whether the glasses would now simply show him what "vitello tonnato" meant if he just looked at it. We had to stop him gently — for a reason that has to do with the Even G2 itself, not a software gap.
Short answer: no, not that way. The Even G2 has no camera — that's a deliberate choice for this device, partly for privacy reasons. Without a camera, no system can capture an image of a sign or a menu and recognise the text in it. What the G2 actually translates is spoken language: it listens through its microphone and shows the translation of what was said on the lenses. For "point at it, see the translation appear," you need a device with a camera — a Ray-Ban Meta, for example, or a camera translation app on your phone.
Why this isn't a software detail — it's a design decision
It's worth explaining briefly why we state this so plainly instead of softening it. The Even G2 belongs to a category of AI glasses deliberately built without a built-in camera. That has a real upside: nobody in the room has to wonder whether they're being secretly filmed or photographed — a point that comes up repeatedly in conversations about wearing them at work or in meetings. But the cost of that decision is just as real: without a camera, there is no image for a text-recognition system (OCR) to read printed text from a sign or a menu. This isn't a feature that's "missing yet" or coming in a future update — it's structurally ruled out by the hardware decision to omit the sensor.
That's exactly why it bothers us when AI-glasses marketing implies "translation" is the same feature everywhere. There are at least two genuinely different tasks, and a camera-free device can only ever solve one of them.
What the Even G2 actually translates
The G2's live translation runs through its built-in microphone: someone speaks, the system recognises the spoken language over a cloud connection, and the translation appears as text on the lenses. That works well for a conversation at a table, at a reception desk, or in a consultation — anywhere language is the medium. It doesn't work for printed text that nobody is reading aloud. If a waiter explains the dish to you, the G2 helps. If you want to read the printed menu yourself, it doesn't — there's no one speaking, just letters on paper.
What actually works for signs, menus and printed text
If visual text translation is what you actually need, there are two honest routes there — both require a camera:
- A smartphone app with camera translation (Google Translate and similar apps): point the camera at the sign or menu and the translated text overlays directly on the camera image. For most situations this is the most pragmatic, already-reliable solution — you have the phone with you anyway.
- Camera-based AI glasses such as the Ray-Ban Meta: these devices have a built-in camera and are designed for exactly this "look and recognise" workflow. In exchange, you give up the camera-free property that's the G2's core advantage — with everything a camera raises about recording ability and discretion around other people.
We're not citing accuracy figures for camera-translation apps or other makers here — those are products outside what we sell, and we're not rating them in detail. The point is the feature category, not a ranking.
The real distinction: two different translation tasks
It helps to separate the two tasks cleanly, because the word "translation" covers both cases even though they have nothing to do with each other technically:
| Task | What it needs | Even G2 (camera-free) |
|---|---|---|
| Someone speaks (conversation, announcement, talk) | Microphone + cloud speech recognition | Yes, core feature |
| Printed text (sign, menu, packaging) | Camera + image recognition (OCR) | No camera |
| On-screen text (foreign-language website, app) | Camera or direct screenshot access | No camera |
Overview for orientation, not a manufacturer specification. For live translation itself — languages, workflow, limits in noise or with dialects — see our translation feature overview.
An obvious workaround, honestly assessed
Some people ask whether they could just read the menu aloud so the G2 "hears" it through the microphone and translates it. In principle that could work in a specific case, since it's the person speaking that reads the text — but then it's the speech case again, not the text case, and reading a menu aloud to yourself in a restaurant is impractical. We're not recommending this; we mention it only to make clear exactly where the technical line sits: at the camera, not the microphone.
So: buy, skip, or combine differently?
Buy the Even G2 if your main need is spoken language — conversations, consultations, travel where someone is talking with you. That's where it's discreet, hands-free, and built precisely for the purpose.
Skip it if your primary reason is "translate signs and menus in front of me" — the G2 structurally cannot do that, regardless of any future software update.
Combine it realistically: the G2 for conversation, a camera-translation app on your phone for printed text. Honestly, that's the setup most travellers already use anyway — just with better discretion on the speech side. If you'd rather have one device for both, look specifically at camera-based models — more on that and the broader feature comparison in our AI glasses database.
Disclosure: AI-Eyewear is the authorised Swiss reseller of Even Realities. We'd rather say plainly "it can't do that" than set an expectation the device can't meet — try the speech translation yourself at the Zürich showroom.