July 2023 Summaries

Live classrooms have become prevalent in education, but network issues can hinder the experience for students in countries like India. WebRTC is commonly used for live video communication, offering low latency but compromising on video quality. HLS, on the other hand, provides high-quality video with higher latency. 100ms combines both technologies to create a seamless live classroom experience by using WebRTC for real-time interaction and HLS for high-quality streaming. This approach allows students to transition between being passive viewers and active participants in near real-time, improving the overall learning experience.

In this tutorial, we learn how to build an audio and video-calling WhatsApp clone using Flutter. The key points of the text are as follows: 1. The calling functionality of WhatsApp is replicated in this demo, but not messaging and authorization. 2. Video/audio calling is accomplished using 100ms' Flutter SDK, while call notifications are instrumented using Firebase. 3. The project starts by cloning the hms-callkit-demo as a starter for this project. 4. A new directory called whatsapp_ui is created inside the lib folder of the Flutter project to create a WhatsApp Home Screen-like interface. 5. A flow is added to add users to the application, and Firebase is used to store the FCM token of the users. 6. The onPressed of the CustomButton on the login_page.dart is updated to store the FCM token and phone number to the Firestore. 7. A chats list is prepared to show on the WhatsAppScreen of the applications using a new file called chats_list.dart. 8. The ChatListScreen() is added on the WhatsAppScreen by modifying the body. 9. Finally, a ChatScreen is added to show individual chats and initiate video calls from there. 10. The app_router.dart is updated to configure Navigation in the application, and the initialRoute in main.dart is updated accordingly. 11. The complete code for this project is available on GitHub.

100ms has introduced new features, Speaker Labelled Transcription and AI-generated Summary, to enhance the analysis of video interactions recordings. The Speaker Labelled Transcription identifies speakers accurately and supports custom vocabulary and multiple file formats. The AI-generated Call Summaries provide a Minutes of the Meeting format with customizable output sections and context settings. These features are currently free in Beta, and future releases will include separate APIs for transcription and summary, standalone transcription and summary without recording, and ability to transcribe multiple roles independently in a single session.