Based on feedback from smartglass users reporting that the app felt "slow," "freezes," and has "confusing controls," we have comprehensively reviewed the experience around small-screen glass devices. We have also made improvements to translate text in AR spaces and fixed issues with offline playback and preloading.

① Smartglass controls have been redesigned

Glass devices differ in screen size and control methods depending on the model. We have standardized the display for each device type in one place. On devices with small screens, we now show thumbnails on a single screen without playing the video.

Thumbnail, single screen: Vuzix M400 series, RealWear Navigator series
Video, split display: dynaEdge XR1, MiRZA, Xreal (as before)

● Added "Description" and "List" to single-screen devices

On single-screen devices, the step text is normally read aloud. However, there are situations where you want to "read it again" or "read it in a place where sound cannot be used." We added Description to the control menu so that step text can be displayed at full screen. Voice commands such as "Description," "description," or "Step text" will also open it.

We also added List (Step list) to the control menu. Previously, users who operated only with the touchpad could not reach the step list without using voice. The control menu order is now: List → Description → Microphone → Video.

● Voice control always starts ON every time you open the app

To prevent distributed devices from being used with voice control turned off, voice control now starts ON every time you open the app. In noisy locations, you can turn it OFF on the spot, but OFF settings are not saved (the next time you open, it will be ON again). RealWear is always ON as before.

● On single-screen devices, we removed Edit, Fix mistakes, Favorites, and Hide header from the header

Because the screen is small and the device is meant for on-site use, these options were just taking up space. "Close" remains available.

● Global zoom has been removed

We have removed the "Global zoom" feature that zoomed the entire screen. Long-pressing for one second could unintentionally trigger it, and once activated, normal focus movement stops, making it difficult to exit and operate the device. We received many complaints about this behavior.

Enlargement and reduction of materials (PDF) remain available through other features and work as before (voice commands "Enlarge" and "Zoom out" also remain unchanged). Zoom on the app side for MiRZA, Xreal, and dynaEdge XR1 also works as before.

● Bug fixes

To address the issue of the app closing when opening the procedure list, we have reduced the number of images loaded in the list. The display that showed four images side by side inside folders is now replaced with a folder icon, and procedure thumbnails are loaded only as they appear on the screen
Fixed the issue where the app became unresponsive for several seconds while the main screen was loading the history of procedures. History is now loaded only when you navigate to "Recent procedures" and does not go back to old formats
Fixed the issue where the focus frame did not align with the actual location that gets pressed
Fixed the issue where the left edge of the focus frame was cut off (in question options, the sheet for opening procedures, and the preload sheet)
Fixed the issue where consecutive voice commands like "play video" would fail to display the video
Fixed the issue where a white bar appeared at the bottom of the screen when opening from the procedure list

Single-screen device control menu (List, Description, Microphone, Video)

② Text in AR spaces is now translated

Previously, even when switching the display language, text placed in AR spaces remained in the original language. The procedure text would be translated, but only the AR space would remain in Japanese.

We have updated the system so that text objects and 3D model names in AR spaces are now displayed in the translated language according to the display language setting. If you switch languages after opening AR, it will switch in real time. Translations also appear when using "View in 3D."

● Please note

In AR editing mode, the original text is displayed (to prevent translated text from being saved as the original)
Characters that are not in the device's font may not display correctly. Please verify on actual hardware before deployment

We also updated the Android app to version 5.0.6 (improved AR movement adhesion and design refinements near the navigation bar).

③ Fixed gaps in offline playback and preloading

We conducted a comprehensive review of offline playback and preloading (Download) based on user feedback and fixed several gaps.

● Bug fixes

Fixed the issue where translated speech audio created during preload would be silent in offline mode
Fixed the issue where "Download complete" was displayed even when saving failed. We now notify you of the number of items that failed to save
Fixed the issue where video saving would stop partway through but still be marked as "Saved," preventing retry on subsequent attempts
Fixed the issue where preload progress appeared to stop midway
Fixed the issue where opening offline would display a previous state of the screen. Once you open it while online, the latest state will be saved
Fixed the issue where thumbnails were cut off in the offline list and the "Downloaded" indicator did not match the actual state
Fixed the issue where preload would continue to run to completion even when the device was set to not allow saving

● Added progress indication

While the data is being prepared for offline use, we now display "Preparing offline data (120/467). Do not close this screen" with a count. If you close the screen during preparation, it will start over, so we inform you of the importance of keeping it open. If preparation does not complete, we notify you rather than claiming success.

● Reduced the amount of data the app stores on the device

We stopped distributing unused images and text in all languages to the device, reducing the amount of stored data by approximately half. This improves the time required for preloading and available device storage.

As a trade-off, if you switch to a language on offline mode that you have never displayed online, it will display in Japanese. If you use multiple languages, please open the screen once in that language before taking the device to the field.

● For devices already preloaded

These fixes become effective when you perform preloading again. For devices used offline, please reconnect and reload in a location with an online connection.

④ AI Search now displays the top result prominently

AI Search is designed to "find one procedure and get back to work," but as more candidates are shown, the selection effort increases. We now display the most relevant result in a large thumbnail card with the full reasoning, and subsequent results are collapsed into "Other matches (N)".

For tasks that require collecting multiple procedures together (linking to skills, building courses), candidates are displayed in full as before.

AI Search results showing the top result prominently displayed with additional matches collapsed

⑤ Other improvements and bug fixes

● Questions

Fixed the "Save answer" button to be fixed at the bottom of the modal. Previously, the action row was left in the middle of the screen when using AR display or full-width display
For photo submission, focus now lands directly on the capture button. Previously, the entire frame including the description was selectable, making it unclear what was being selected
Added the ability to select and delete photos one at a time after they have been captured
Fixed the issue where after going to the end of the questions and returning to the beginning, the question text remained off-screen and unreadable

● Playback and text-to-speech

Fixed the issue where subtitle narration became distorted and difficult to understand
Revised how subtitles are shortened when they do not fit in the frame, making them easier to follow on screen