Patch Notes — Version 1.4.0 Fixed known bugs. gemini-3.7-flash added

Patch Notes — Version 1.3.9 Please check manual for details. New Features: 1.AI reasoning level, Platform Models now use adjustable Reasoning Levels instead of separate Thinking and Non-Thinking variants. Available levels: Auto, Off, Low, Medium, High, and Max. Higher levels may increase AI credit usage and response time. Auto: Lets the AI automatically choose the appropriate reasoning level., Off: Disables reasoning (equivalent to the previous Non-Thinking)., 2. AI Configuration on the 7-Cal, Reasoning Level, Online Search, AI Memory, and Manage Files to AI can now be configured directly on the 7-Cal in Setting → AI Config. 3.LaTeX settings are now available in Setting → Advanced → LaTeX., 4. New 500 AI Credits Package, A new 500 AI Credits package is now available on our website for USD $25, with 365-day validity. This provides a more convenient option for users who need larger amounts of AI credits without purchasing multiple 100 AI Credits packages. 5. New AI Models, Added support for Kimi K3, GPT-5.6 Sol, Claude Opus 5 and Claude Sonnet 5.

New Features Added Keyword Search on the device Users can now use keyword search in AI Chat, Chatroom, Online Files, and Offline Files. Added Upload File to AI in the Backend Panel Users can now upload PDF files directly for the AI model to reference. This feature can be managed in Backend Panel → AI Models. Because uploaded files may consume more AI credits than normal AI chat, we recommend uploading and activating a PDF only when you specifically need to provide that file to the AI. After uploading, select Active for the PDF you want the AI model to use. Users can upload up to 5 PDF files, with a total limit of 100 pages or 10 MB. Improvements & Fixes Fixed several bugs.

Please check the manual for details. New Features 1. Added Network Speed Test under Network Settings. Users can now test the current Wi-Fi or 4G network speed directly on the device, with results shown in KB/s. 2. Added online editing for offline files. After connecting to the device hotspot, users can now view, edit, and save text files directly from the web page. 3. Improved message preview behavior. When users exit a preview and open it again, the device now returns to the previous reading position. 4. Added a LaTeX Preview button in Chatroom, allowing users to check LaTeX syntax and preview how it will appear on the device before sending. Improvements & Fixes 1. Improved 4G connection reliability. 2. Fixed an issue where right-to-left languages, such as Arabic and Hebrew, could display multi-paragraph text in the wrong order. 3. Fixed a brief keyboard flashing issue after sending messages with the virtual keyboard in AI Chat. 4. Rebuilt LaTeX rendering on the device for clearer formula display and fixed several rendering bugs. 5. AI response history on the website now shows which AI model was used. 6. Custom Models now show clearer error codes when testing fails. Saved Custom Models can now be edited.

Patch Notes — New Features & improves Please check the manual for details. 1. Added GPT-5.5 and GPT-5.4 Non-Thinking Models 2. Backend Panel Additions: You can now manually control the following features in the Backend Panel: · AI Chat Memory. Important Notes About AI Memory: AI Memory may slow replies and use more credits because extra context is sent to the AI. Memory will only clear when press DEL or when the device is powered off. AI Memory is shared across all models, including Custom Models. · Online Search. Important Notes About Online Search: When enabled, Online Search will automatically apply to all supported platform models and Custom Models. If your Custom Model returns errors after enabling Online Search, please disable the feature. 3. Removed Code Sandbox The previously added Code Sandbox feature has been removed. Very few users were actively using the feature. The feature increased AI response time. It has been removed to improve overall AI response speed and performance.

Patch Notes — Version 1.3.6 Please check manual for details. New Features 1. Added Anti-Shake settings for the camera. Multiple anti-shake levels are now available for clearer handheld photos. Note：Higher Anti-Shake levels may make photos clearer when handheld, but performance may become worse in dim or low-light environments. 2. Added “R” reply indicator. 3. Added Startup Network option. You can now choose WiFi or 4G as the default startup network in 4G Setting option 3. 4. Added mobile support for the Chatroom webpage. The Chatroom webpage can now be used more easily on phones. 5. Added on-device deletion for Offline Files. Press DEL to delete. 6. Added WiFi switching logic. Users can now switch from the current WiFi network to another saved WiFi network. Improvements & Fixes 1. Improved camera stability by optimizing camera parameters. 2. Improved camera capture speed and overall response. 3. Increased Offline Files limit to up to 100 files. 4. Fixed several LaTeX rendering bugs. 5. Improved RTL logic when LaTeX display is enabled.

Improvements & Fixes Fixed an issue where some incorrectly formatted .txt files could not be uploaded to Offline Files Improved RTL (right-to-left) display formatting for Arabic and Hebrew Fixed issues where certain LaTeX formats failed to render properly

Improvements & Fixes Added some protection layers for the 4G version to avoid some unusual crashes when connecting to a 4G network.

Patch Notes — Version 1.3.2 Improvements & Fixes Fixed an issue where the chatroom does not show a long LaTeX message Some minor touch-up on the 4G function to help with the connection. Patch Notes — Version 1.3.3 Added Claude Opus 4.7 Fixed several LaTeX rendering bugs on the device Added Code Interpreter sandbox support to all thinking models This allows the model to run Python code safely in a sandbox for math, debugging, charts, and other logic-heavy tasks To use it, follow the shared prompt in ⁠https://discord.com/channels/1396394690860814417/1494626755514863616

Added Grok 4.20 thinking model Fixed a bug in Calculator Mode Updated the default camera settings to have better anti-shake

Patch Notes — v1.3.0 Please check manual for details. New Features: 1. Auto Scroll: Auto Scroll will now automatically pause, scroll horizontally to the right until the formula has been fully traversed, and then resume vertical scrolling. 2. HoldScroll/Auto Scroll update:The repeat speed for long-pressing directional keys is now separated from the normal Auto Scroll speed. 3. Expanded Auto Scroll Range: The Auto Scroll interval range has been expanded: Previous range: 100–2000 ms New range: 100–20000 ms (1–20 seconds) Improvements & Fixes: 1. Fixed an issue where some users were unable to delete saved Wi-Fi information from the device. 2. Optimized Wi-Fi connection parameters Improved Wi-Fi connection compatibility to address connection issues reported by some users.

Please check manual for details. New Features 1.Auto Scroll Re-enabled: Auto Scroll has been brought back across supported reading and message views. 2. Bottom-to-Top Loop Navigation: When reaching the end of content, users can now jump directly back to the top. The End hint is now always shown at the bottom of the content. The hint now displays as: [End, press ↓ to top] Pressing Down at the end will now immediately jump back to the beginning Improvements & Fixes: Fixed WDT crash caused by Wi-Fi status rendering, Fixed formula loading getting stuck indefinitely, Fixed shared horizontal scroll offset for wide formulas, Fixed Offline files search functions

Please check the manual for details. New Features 1. Keyboard Layer Quick Switch You can now press the 0 key to quickly switch between keyboard layers. 2. AI Chat Message Queue You can send multiple AI messages without waiting for previous replies. All messages will be processed and displayed in sequence. 3. Offline Files Keyword Search Search for specific keywords within offline documents.

Please check the manual for details. New Features 1. LaTeX Display Support (Beta) LaTeX rendering is now supported. You can enable this feature in the Backend Panel → AI Model page. Once enabled, LaTeX formatting will be displayed in Online Files, AI Chat, and Chatroom. Please note that this is currently a beta feature. 2. Offline File Storage Added support for offline file storage. 3. Hebrew RTL Support Added Hebrew right-to-left (RTL) display support. 4. Stealth Mode for Camera Improvements & Fixes Fixed several known issues. General system stability improvements. AI Chat History Clear Function

• GPT-5.3 • GPT-5.4-Think • GPT-5.4-Pro

1. Calculation Bug Fixes Fixed a calculation bug in Calculator Mode that could cause incorrect results under certain conditions. 2. Custom Screen Brightness Control Users can now manually adjust the screen brightness. 3. Black Images Bug - Automatic Re-shoot Algorithm Fixed an issue where the camera could capture black images. Introduced a new re-shoot mechanism. When the system detects a black image, the device will automatically re-capture the photo. 4. Brick Mode When Brick Mode is enabled: • All functions except Calculator Mode are forcibly disabled • Wi-Fi, BLE, and other non-calculator features are turned off, and no wireless signals are transmitted. 5. Network Lock (4G + LTE Version Only) In a 4G network connection, the device will now connect exclusively to 4G LTE networks, ignoring 3G and other legacy network types. Improvements 1. Password System Optimization: Now supports custom passwords with 1–8 keys. 2. Keyboard Input Enhancements: Improved keyboard input experience, added additional selectable characters and extended mathematical symbol input keyboards

This update focuses on improving 4G stability, refining calculator mode, and adding several highly-requested features for a smoother overall experience. Please check manual for details. 1. 4G Improvements • Resolved 4G connection issues, improving registration speed and overall stability. • Added a detailed 4G connection guide in manual to help users troubleshoot and configure their SIM cards more easily. 2. Calculator Mode • Fixed all known calculator mode bugs reported up to this version. • Updated the calculator UI to match the traditional 82MS layout with complete accuracy. 3. AI Chat Enhancements Added a manual “Clear AI Memory” option to reset the last 5 conversation rounds at any time (Press DEL key in ai chat). 4. Improved reading experience: Auto Scroll: Hold the Down key to scroll continuously while viewing documents or long AI responses (Adjustable in Advance settings). 5. Camera Improvements • Added manual photo-quality adjustment in the advanced settings. This setting controls image compression: lower values apply less compression, resulting in higher image quality and larger file sizes. • Added manual resolution control in the advanced settings. Lower resolution provides better anti-shake (stabilization) for handwriting capture. 6. Connectivity & Convenience • Added Wi-Fi memory support — the device now remembers up to 10 previously connected networks, prioritizes connecting to the last used network. Saved Wi-Fi networks are indicated by a small dot; highlight the network and press DEL to remove the saved information. • The device will also detect Wi-Fi networks with hidden SSID enabled and display them as Hidden Network. To connect to a hidden network, you must manually enter both the Wi-Fi name and the password. • Added an on-device credit checker in setting → 6.credits, allowing users to view their remaining AI credits directly from the device.

GPT 5.2, GPT 5.2-THINK, and GPT 5.2-PRO have been added. GPT 5.2-PRO is OpenAI's best AI model and also the most expensive one. 5.2-THINK is already enough for most complex tasks.

Important Notice: This is an experimental update. We DO NOT recommend upgrading if you rely on your device heavily at the moment. Please remain on your currently stable version unless you specifically wish to try the new features. Camera Improvements Further refined the camera algorithm. Even under extremely humid, low-light environments, photos now remain clear, bright, and easy to read. Wi-Fi Enhancements 1. Improved compatibility with Wi-Fi networks that require both username and password authentication. 2. Significantly increased Wi-Fi connection speed across various routers. 3. Fixed an issue where certain Wi-Fi networks could connect successfully but failed to reach the server afterward. Network Stability Updates 1. Optimized device spoofing parameters to reduce cases where public networks reject the connection. 2. Improved overall stability when using 4G mode.

Follow-up hotfix for 4G mode Some fixes in calculator mode

This hotfix release addresses several urgent issues reported by users and introduces early improvements for Calculator Mode. Fixes 1. Fixed a reboot issue affecting the 4G version of the device, improving system stability under mobile network usage. 2. Optimized Wi-Fi connection speed and general connection performance. 3. Improved stability when connecting to non-mobile-hotspot Wi-Fi networks, especially router-based networks. New (Beta) Added beta support for Wi-Fi networks that require a username and password (WPA-Enterprise). This feature is in early testing and may not work with all network configurations. Calculator Mode Implemented initial fixes for several calculation bugs in Calculator Mode. As this is a hotfix, the calculator corrections are not fully completed and will continue in the next update.

1. 4G Network Improvements Complete rebuild of the 4G connection framework, resulting in faster network registration, higher stability, and wider carrier compatibility. 4G-only battery life improved to an estimated 8–10 hours through system-level power optimizations. 4G network speed improved by approximately 2×, depending on signal conditions. • Added real-time 4G connection step display, showing each stage of the attachment process. • Added APN settings for users who need custom operator configuration. • Added IMEI viewing to assist with carrier SIM activation. ⸻ 2. Connectivity & Status Bar • Added WiFi and 4G status bar icons, both with real-time signal strength indicators. • OTA updates are now supported on both WiFi and 4G (previously WiFi-only). ⸻ 3. Camera Enhancements • Added Manual Focus option, allowing users to switch between Auto Focus and Manual Focus. Manual Focus significantly reduces capture time and is ideal for repeated captures at a fixed distance. • Improved Auto Focus speed and stability. • Both focus modes are selectable in Advanced Settings. ⸻ 4. AI Mode & UI Improvements • Added AI Mode Resume, allowing 7-Cal to return to the previous AI screen upon re-entry. • Updated upload status UI in AI Chat - now displayed at the top of the screen for better clarity. • Improved model selection and OTA update user interface, with clearer prompts and progress indicators (based on user feedback). • Added customizable keyboard sensitivity settings. ⸻ 5. Fixes & System Stability • Fixed inaccurate battery percentage reporting for more consistent readings. • Improved 4G error logs with more detailed and actionable diagnostic information. • General system optimizations to enhance overall speed, responsiveness, and long-term stability.

1. 4G Mode Hotfix Fixed an issue affecting 4G connectivity stability in certain regions. 2. Camera Quality Improvements (Early Testing Feature) Added an intelligent brightness-boosting algorithm for low-light environments to compensate for the lack of a hardware flash. • In dim lighting, the camera now automatically increases exposure and adjusts focus parameters. Low-light captures may take 4–20 seconds to ensure proper focus and clarity. • This feature is in early testing and will continue to be refined in upcoming versions. • Normal-light photography is unaffected. 3. Improved Battery Level Detection Optimized the algorithm for more accurate battery percentage reporting. 4. Enhanced Key Sensitivity Adjusted button response for a smoother and more consistent typing experience.

1. Unified AI Mode AI Chat and AI Camera have been fully merged into a single AI Mode. Users can now type messages and take photos seamlessly in the same interface, greatly simplifying the workflow. 2. New Feature: Chatroom Added Chatroom functionality, supporting up to 10 participants per room. Both text and photo messages are supported, using the same interaction flow as AI Mode. Note: The Chatroom interface in the backend panel is desktop-only and cannot be accessed on mobile devices. 3. Photo Limit Increased Each chat message now supports up to 10 photos, improving handling for multi-step problems and batch photo submissions. 4. WiFi Transfer Speed Improved Data transfer speed over WiFi has been improved by approximately 2×, resulting in smoother uploads and improved responsiveness. 5. Minor Bug Fixes Several issues have been addressed to enhance stability and overall user experience. More details can be found in the latest manual. If you prefer maximum stability, you may choose to skip this update for now.

Message history memory has added for all AI models, it will remember the most recent 5 round of chat history and images, memory will be reset when you switch models or disconnected the device. This is a server update, no need to update device.

1.Resolved double-click issue affecting certain users when pressing keys . 2.Fixed “Upload Failed” error that occurred when captured images exceeded the allowed size limit. 3.Improve upload speed under WiFi New (Beta) Introduced Chatroom (Beta) — an experimental new feature that enables live chat directly from the device. ⚠ This feature is still unstable and may experience occasional crashes or connectivity issues.

We recommend updating to Version 1.1.6 for a more natural and user-friendly photo-taking experience. This update enhances the photo capture experience with a more ergonomic design: 1. Updated the camera key layout — previously, taking a photo required pressing the “=” key. Now, users can press any key to capture a photo. Note: “Any key” does not include the AC key, left return key, or MODE key. This change improves ergonomic comfort, as the “=” key’s lower position made one-handed photo capture less convenient for many users.

We recommend all users update to this version to ensure system stability during network transitions and firmware updates. This hotfix addresses critical stability issues related to network switching and OTA updates: 1. Fixed a crash that could occur when switching from 4G to Wi-Fi while connected to the 4G network. 2. Added a new safety reminder: when performing OTA updates in 4G mode, the device will now prompt users to switch to Wi-Fi for a more stable and faster update process. 3. Fixed other minor issues

This update focuses on improving overall 4G connection stability and reliability. 1. Enhanced 4G network stability for smoother connectivity. 2. Added a recovery mechanism for rare cases where the 4G module may freeze during connection attempts - pressing the AC key will now safely return the device to normal operation. 3. Fixed several minor bugs related to 4G network registration and data transmission. We recommend all 4G users update to this version for a more stable and responsive experience.

Summary: This version was jointly developed with our 4G IoT SIM card provider to further address network registration issues experienced by some users. Details: Improved network compatibility and registration reliability for 4G models in specific regions. Enhanced carrier profile handling for smoother activation and connection stability. Verified in collaboration with LTE provider engineering teams. Scope: Applies only to 4G models using the our SIM card. Wi-Fi version users are not affected. Type: Hotfix Recommendation: 4G users are advised to update to v1.1.2 to ensure optimal connectivity across all supported regions.

Summary: This hotfix resolves an issue affecting some countries and regions where certain LTE carrier configurations prevented 4G SIM cards from successfully registering on local networks. Details: Fixed: Network registration failure caused by LTE operator setting inconsistencies in specific regions. Scope: Applies only to 4G versions of 7-Cal. Note: Wi-Fi version users are not affected by this update. Type: Hotfix series update Recommendation: 4G users in affected regions are advised to update to v1.1.1 as soon as possible for improved connectivity stability.

Online search feature has added to GPT-5-THINK model, you can add a custom prompt asking it to do a online search before answering, please aware other models currently cant do online search, if you ask it to do, it will cause error, only let ai do online search when its using gpt-5-think. Online search for other model will be added in future.

Important: This update is recommended for all users. Fixes: 1.Freeze when Wi-Fi model tries to use 4G On Wi-Fi-only hardware, attempting to enable 4G could cause the device to hang. Now: after a short stall, the system automatically aborts the 4G attempt or performs a safe reboot to recover. 2.Incomplete characters on screen Fixed rendering issues where some special symbols and certain languages were not fully displayed. Improvements: Faster OTA updates: Optimized download/installation pipeline to reduce update time. Important Notes (Wi-Fi Model): The Wi-Fi-only version does not support 4G. We do not recommend turning on the 4G option in Settings on the Wi-Fi model. Doing so may still trigger a recovery reboot or brief unresponsiveness while the system exits the unsupported operation. How to Update: Update via Settings → Update (OTA). Ensure stable power during the process.

You can now try the new Custom Prompt feature on your user backend, under the AI model section! This allows you to set your own default AI behavior, for example: Fix the reply language (helpful when AI misdetects due to math symbols or mixed text). Instruct AI to only answer the first question in a captured photo. While you can still type these instructions manually after each capture on the calculator keyboard, the Custom Prompt lets you save some typing, improving efficiency and consistency in your workflow.

Important: We strongly recommend all users update to v1.0.8 immediately after receiving the device. What’s New 1) Smart Compression for Document Photos New codec tuned for paper/printed text. Typical file size ~150 KB per image while maintaining text sharpness for OCR and reading. Trade-off: Color fidelity is slightly less accurate than previous versions. Benefit: Much faster uploads over the internet (better experience on slower or metered connections). 2) Enhanced Autofocus with Anti-Shake Filtering New stabilized AF pipeline designed for handheld capture. Filters minor hand tremor to preserve text clarity, reducing the blur seen in prior releases. The advanced AF routine takes about ~4 seconds end-to-end (trigger → focus → anti-shake → save). How to Capture the Sharpest Text (Recommended Use): Keep ~20 cm (≈8 in) between the device and the document. This distance helps the camera lock focus reliably on printed text. If you must shoot closer than 20 cm, keep the device perfectly parallel to the page. Tilting at close range reduces focus margin and can soften edges. Audible focus lock cue: When focus is acquired, you may hear a crisp mechanical “click” from the lens module . This indicates the device has locked focus and is ready to capture. Notes & Trade-offs The compression update prioritizes legible text and smaller files over perfect color reproduction. AF duration may vary slightly with lighting/contrast; ensure good, even lighting for best results.