Skip to main content

2026

04-13​

Introduced SDK Versioning (2.0 / 3.0)​

To support the video intercom refactor, we re-architected the camera-centric pipeline and clearly separated components such as capture, ivcore, display, facial, and intercom, so the same camera stream can be used concurrently for multiple scenarios including intercom, display, and recognition. Because the new and legacy component stacks differ at both architecture and interface levels, they are no longer backward compatible.
To make version management and project selection clearer, we introduced the SDK version concept: SDK 2.0 includes all versions in the legacy component stack, while SDK 3.0 includes all versions in the new component stack.

04-02​

Video Intercom Support for Face Recognition Devices​

Face recognition devices now support video intercom, enabling visitors to initiate a door call and allowing users to answer from indoor units or mobile apps with real-time video and two-way audio. After confirming the visitor identity during the call, users can complete remote door release as part of the same workflow, with support for private server deployment.
For detailed architecture and call flow, refer to the Video Intercom System Overview documentation.

03-06​

Vibe Coding (AI-Assisted Development) Documentation Added​

New Vibe Coding section in the docs: an overview of AI-assisted DejaOS development, reusable prompts for Cursor/Claude Code to generate DejaOS code correctly, and a UI Manager reference with full sample code. Helps developers and AI tools stay aligned with DejaOS conventions and ship features faster.

02-20​

Apps Module Launched​

We have launched the Apps & Samples section, which showcases examples ranging from feature combinations to full scenarios. All content in this section is fully AI-generated, including complete screenshots and full business logic, so developers can reference or extend the samples directly.

01-19​

VSCode DejaOS Plugin v1.0.39 Released​

  • Enhanced automatic component version selection - click a button to automatically select the latest version for a single component or all components at once.
  • Improved component search functionality with faster and more accurate results.
  • Various UI/UX improvements for better development workflow.

01-10​

AI-Generated Smart Locker Application Released​

We are excited to announce the release of a fully functional Smart Locker System - an application entirely generated through AI-assisted development!

Highlights​

  • 100% AI-Generated: This complete embedded application was created through 100+ conversations with AI assistants, demonstrating the power of AI in full-stack embedded development.
  • Production-Ready: Unlike demo examples, this is a fully functional application ready for deployment with minimal customization.
  • Offline Operation: The current release is an offline version that operates completely independently without any network connection.

Key Features​

  • User Functions: Store items with 6-digit password, retrieve items, real-time locker availability display
  • Admin Functions: Locker group management, batch cabinet control, access records, time settings, password management
  • Multi-Worker Architecture: Separated UI, business logic, and hardware communication for optimal performance
  • Hardware Integration: RS-485 communication with lock control boards, PWM buzzer feedback

Coming Soon​

An online version with cloud connectivity and facial recognition support is planned for future release.

For detailed documentation and source code, please visit our Smart Locker System (Offline Version) documentation.