INMO GO3 smart glasses add Gemini and a display, with a camera you can cover

INMO GO3 smart glasses combine an in-lens display with advertised Gemini support, while a physical cover lets wearers block the camera. That combination addresses a practical conflict: glasses need useful functions when their cameras are covered, especially around people who do not want to be photographed.

The product offers a different balance from glasses built primarily around taking pictures or delivering spoken answers. Its display creates room for readable assistance, while the cover offers a visible way to limit image capture. Neither feature establishes that the entire device is private.

Meta Ray-Ban Display already provides a reference for glasses with visual output, so INMO is not introducing the category’s first screen. The more interesting question is whether its combination makes sense throughout an ordinary day. A display, camera, and AI assistant must earn their place together, including when the wearer deliberately disables visual input.

INMO GO3 smart glasses bring the display and camera together

The GO3’s central proposition is useful visual assistance with a camera that the wearer can physically cover. That is a more specific claim than replacing a phone or making every conversation smarter.

The chronology matters. INMO’s GO3 launch announcement, dated April 2, 2026, described an upcoming Kickstarter campaign following an earlier CES appearance. October coverage therefore should not be confused with the product’s original announcement.

That announcement already described the camera cover and replaceable battery approach. These are established parts of INMO’s product pitch, rather than features that first appeared with the latest headline. The distinction helps readers separate renewed attention from a newly announced capability.

INMO currently advertises Gemini-powered assistance, alongside ChatGPT support, translation, navigation, meeting transcription, and a teleprompter that displays a prepared script. These are company descriptions of available functions, rather than findings from an independent test conducted for this article.

The attraction is easy to understand. A short answer shown near the wearer’s line of sight serves a different purpose from an answer spoken aloud. Text lets someone inspect wording, reread a phrase, or glance at information without broadcasting the response.

Consider a hypothetical conversation where a visitor needs help following an unfamiliar language. Readable assistance offers a way to check a phrase while staying engaged with the other person. That scenario explains the product’s appeal without establishing its translation accuracy.

The camera adds another input path. It gives an assistant a possible route to interpreting something in front of the wearer. Covering it deliberately removes that route, creating a useful boundary between assistance based on speech and assistance based on images.

This distinction is central to evaluating the device. The question is not simply how many functions appear in its menu. It is which functions remain useful with the lens blocked, and how clearly the software communicates that boundary.

A good experience would make the choice understandable before the wearer starts a task. If an action requires an image, the interface should say so. If the action only needs audio or prepared text, the camera should not become an unnecessary dependency.

That is an evaluation standard, not a verified description of every GO3 workflow. It provides a concrete way to assess the product beyond a demonstration. Useful glasses should respect the wearer’s chosen inputs instead of constantly asking for broader access.

A covered camera changes what people must trust

A physical cover offers a narrow, understandable assurance: it blocks the camera’s view when properly fitted. It does not establish what the microphones, companion software, or remote services are doing.

That narrowness is also its strength. People standing nearby cannot inspect a settings screen inside someone else’s glasses. A visible obstruction gives them something external to evaluate, provided they can recognize it and see that it covers the lens.

INMO’s GO3 hardware specifications list a binocular green display, meaning visual output for both eyes, with a 640 × 480 resolution. The same page includes camera covers among the package contents.

Those details support a straightforward reading of the design: the display provides a reason to keep wearing the glasses after blocking the camera. Whether that reason is compelling depends on software quality, readability, and the tasks someone actually performs.

The main tension is between persistent assistance and understandable limits on capture. A wearable assistant benefits from convenient access to information. People around its wearer benefit from knowing which information the device can collect at a particular moment.

Research supports taking that social problem seriously. A camera-glasses privacy study, published at CHI 2024, followed 15 participants over two weeks. It found that wearers experienced responsibility for bystander privacy and wanted clearer indicators.

That study examined an earlier generation of camera glasses, not the GO3. It does not validate INMO’s design or measure reactions to its cover. It does show why uncertainty about recording affects the person wearing the device as well as nearby people.

The practical implication is that privacy controls have an audience beyond the owner. A control hidden inside an application can satisfy the wearer’s preference while leaving everyone else uncertain. A visible control has the potential to communicate across that gap.

Potential does not guarantee understanding. Someone unfamiliar with the product might not identify the covered component as a camera. A small cover might also be difficult to notice across a table, under poor lighting, or during a brief interaction.

There is another boundary to explain. Blocking an image sensor says nothing about audio capture. A person seeing the cover should not have to infer that every recording function has stopped, particularly when the product also promotes conversation-related assistance.

For workplace use, this creates a practical purchasing question. Buyers need clear descriptions of camera status, microphone status, and stored information. They also need a way to explain those states without requiring every colleague to understand the glasses’ interface.

The cover therefore deserves credit as a concrete design choice. The larger privacy claim requires separate evidence about communication, recording behavior, and data handling. Treating those as distinct questions makes the product easier to assess fairly.

The display creates a role beyond taking pictures

The screen matters because it gives the glasses a purpose during moments when image capture is inappropriate or unnecessary. That purpose must be strong enough to justify wearing the device.

There is already meaningful competition here. Meta’s display-glasses announcement introduced Meta Ray-Ban Display in September 2025, with visual assistance and wristband-based interaction. INMO therefore competes in an existing category of wearable visual interfaces.

That comparison establishes context without establishing a winner. A screen’s presence does not settle readability, comfort, response quality, or control accuracy. Those qualities require comparable tasks and conditions, ideally using current software on both products.

For INMO, the most interesting comparison is how well the product works after the camera is covered. That choice shifts attention toward readable responses and other functions that do not inherently require a view of the surroundings.

A hypothetical presentation illustrates the distinction. A speaker who wants access to prepared notes does not necessarily want a camera aimed at the audience. Display-based assistance and image capture serve different purposes in that setting.

The same reasoning applies to private reading. A wearer might want to inspect a short piece of information without speaking it aloud. The value comes from where the answer appears, rather than from the device collecting more information.

This creates an editorial test for smart-glasses demonstrations. Show what happens when the camera stays covered throughout the task. If the useful experience survives, the device has a stronger claim to being everyday eyewear with assistance.

If useful functions repeatedly stop or become confusing, the cover becomes harder to keep in place. That would weaken the design’s practical benefit, even if the accessory itself blocks the lens perfectly. Hardware boundaries need software that accommodates them.

The screen also introduces its own compromises. Information near the eyes competes with the person or environment in front of the wearer. A readable answer is only helpful when its timing and presentation fit the situation.

A short prompt during a pause differs from text arriving while someone crosses a busy street. Similarly, a concise response differs from a paragraph that requires sustained attention. These are interface questions, not conclusions about GO3 performance.

Developers should pay attention to those distinctions. A wearable assistant needs decisions about when to show information, how much to show, and how to dismiss it. Moving a phone interface closer to the eye does not resolve those decisions.

Knowledge workers have a related concern: whether information remains useful after the moment passes. An answer that disappears is different from a record someone can review later. Durable knowledge management requires organization and retrieval beyond the display itself.

That does not make glasses an unsuitable capture tool. It means their value should be assessed across the full workflow, from obtaining information to checking and reusing it. Convenient capture is one stage of that process.

The competitive pressure is therefore concrete. Display-glasses makers must show repeatable benefits during normal tasks, including tasks performed with restricted inputs. The GO3’s cover makes that requirement particularly visible because it invites the wearer to choose a limit.

Gemini support needs a clearer definition than a logo

Advertised Gemini support establishes an AI feature claim, but it does not independently establish how that feature performs or handles information. Buyers need the details behind the integration.

Gemini is Google’s family of AI models and assistant products. Naming it tells readers something about the advertised source of assistance. It does not, by itself, specify which model is used, which tasks invoke it, or how the connection operates.

Those distinctions matter because an assistant’s name and a device’s overall software platform are different things. Support for an AI service should not be treated as proof of a particular operating system, developer environment, or partnership arrangement.

The relevant questions begin with the actual interaction. What does the wearer request? What information accompanies that request? Where does the response appear? Which parts of the experience depend on a connected phone or an available network?

The next questions concern failure. An assistant should communicate when it lacks an image, cannot understand speech, or cannot complete a request. A fluent answer is not enough if the wearer cannot tell whether it reflects the available evidence.

This is particularly important for visual assistance. A useful system needs to distinguish between an unclear image and an identifiable object. It should also make uncertainty legible within the limited attention available during an ordinary interaction.

Independent reporting provides a reason to avoid treating advertised functions as universally verified. In a September 29 review, ITmedia journalist Takahiro Koyama described limitations encountered while testing an earlier iOS application environment.

His GO3 camera testing reported that advertised image-related AI functions were unavailable in his setup as of late August. The manufacturer told him those functions should operate normally, leaving a specific discrepancy requiring further testing.

That finding has a clear scope. It concerns one reviewer’s environment at an earlier date. It does not establish that the same problem affects every owner, every phone, or the software available in October.

It nevertheless identifies the right standard for future coverage. Reviews should record the application version, glasses software version, phone platform, and exact task. Without that context, positive and negative reports can describe different products in practice.

For buyers, this makes task-based verification more useful than a list of supported AI brands. Someone interested in understanding printed text should see that exact action performed. Someone interested in spoken assistance should evaluate a representative conversation.

The same principle applies to meeting-related features. A polished summary does not establish whether the underlying transcript captured names, decisions, or qualifications correctly. People who depend on those records need to inspect the source material and correct errors.

No specific GO3 accuracy rate is established here. That absence should remain visible rather than being filled with assumptions about the underlying model. A familiar AI brand cannot substitute for evidence about the complete device experience.

There is also a distinction between capture permission and processing permission. Agreeing to a photograph does not necessarily answer where that photograph will be analyzed or stored. Product documentation should make those stages understandable to the wearer.

Enterprise evaluation adds questions about access, deletion, and retention, meaning how long information remains stored. These are purchasing criteria rather than verified deficiencies in the GO3. The point is to ask them before sensitive material enters a workflow.

A clear explanation would help INMO as well as its customers. Precise descriptions reduce confusion about what Gemini support includes and what remains outside it. They also give reviewers a stable set of claims to test.

The result should be a smaller, better-defined promise. Readers need to know which tasks work reliably under stated conditions. That is more useful than assuming the presence of a recognizable assistant makes every advertised scenario equally mature.

Replaceable batteries address interruption, with another tradeoff

A replaceable battery changes how the wearer handles a depleted device; it does not remove the need to manage power. The benefit is continuity, provided the spare is ready.

INMO advertises a five-second battery swap and up to three hours of continuous translation use. Those are manufacturer claims, and the continuous-use figure should remain distinct from broader language about wearing the glasses throughout the day.

The distinction changes the practical meaning of endurance. A person might wear glasses for a long period while using their electronic functions intermittently. Another person might need sustained assistance during a conversation-heavy task. Their experiences will differ.

A removable battery offers a way to recover without waiting for the glasses themselves to recharge. In principle, that makes the interruption easier to schedule around a break. Whether it feels convenient depends on the actual handling and recovery process.

That process includes more than removing and inserting a component. The wearer needs to know when power is low, locate the spare, complete the change, and resume the interrupted activity. Each step contributes to the experience.

The software response also matters. Does the device return to the relevant screen? Does an active session survive? Are unsaved notes preserved? These are useful review questions, and no universal answers should be inferred from the advertised swap time.

A hypothetical interpreter-assisted meeting shows why recovery deserves attention. Replacing a battery between sessions is different from losing assistance during an important exchange. The consequence depends on timing as much as on the physical replacement.

A spare also becomes another object to manage. Its charge state needs to be understandable, and carrying it must fit the wearer’s routine. The convenience of a replaceable design depends partly on preparation before leaving home.

This does not diminish the approach. It identifies the tradeoff accurately. The user gains a way to restore operation while taking responsibility for another charged component. Some routines will suit that exchange better than others.

Comfort deserves the same practical treatment. A specification sheet cannot establish how glasses feel on a particular person after sustained wear. Fit, balance, prescription needs, and control access all affect whether the electronic functions remain worth using.

The strongest evaluation would combine those factors in a normal task. Wear the glasses, use the relevant assistance, replace the battery when needed, and check what was interrupted. That sequence would reveal more than an isolated battery demonstration.

For workplace buyers, the lesson is to measure completed activity rather than headline endurance. A device that recovers predictably can be useful despite limited continuous operation. A device with longer advertised endurance can still frustrate users if its workflow is unreliable.

The GO3’s battery design therefore belongs in the same analysis as its camera cover. Both provide a physical choice that changes everyday use. Their value depends on whether the surrounding software and carrying routine make that choice easy to sustain.

Three signals will show how the GO3 proposition holds up

The next useful evidence should show current task reliability, understandable privacy boundaries, and continuity during extended use. Those signals would make the product easier to judge than another feature announcement.

First, watch for repeatable tests using identified application and glasses software versions. The earlier ITmedia finding makes image-related assistance an especially useful checkpoint. A current demonstration should show the complete request and result under clearly described conditions.

Repeated success across supported phone environments would strengthen confidence in the advertised experience. Persistent failures under those same conditions would weaken it. Either result would be more informative than reports that omit the software and task details.

Second, watch how the camera cover works as a communication tool. The relevant question is whether nearby people understand what it blocks. Reviewers should also examine whether the interface keeps camera status and audio status clearly distinguishable.

Evidence of consistent understanding would strengthen the privacy argument. Confusion about continuing audio capture would limit it. A physical barrier has a specific job, and good communication should keep its meaning equally specific.

Third, watch for extended-use accounts that include an actual battery change. Useful reports should describe what happens to an active task and whether the wearer returns to it predictably. The handling and recovery deserve attention alongside the replacement itself.

Reliable recovery would support the case for the swappable design. Lost context, unclear charge status, or repeated setup would reduce its benefit. These outcomes would show whether the hardware choice improves the full experience.

For now, INMO GO3 smart glasses present a coherent idea: readable assistance can remain valuable when a camera is deliberately covered. The verified design elements support examining that idea seriously. They do not justify declaring the broader privacy or usability problem solved.

If you are following this category, start with the task you want glasses to improve. Ask to see it performed with the camera covered, then with the camera available. Check which information the device needs, where the result appears, and what happens when the request fails. For work-related use, inspect what remains afterward and how you would correct or retrieve it. Finally, look for a review that repeats the task after ordinary wear and a battery change. That sequence turns an appealing hardware combination into an assessable workflow. Would the glasses still earn a place in your day when you choose to limit what they can capture?