v-modal/vmodal_sdk_smartglass

What it solves

It simplifies the complex process of streaming live, first-person video from Android-based smart glasses to a cloud ingest service. It removes the "media plumbing" burden from developers, handling the difficult tasks of hardware encoding, secure transport, and network recovery so they can focus on the user experience.

How it works

The SDK uses a hybrid architecture combining a Kotlin client for Android-native management and a Rust core for high-performance media handling.

1. Kotlin Client: Manages the public API, camera coordination, and the lifecycle of the capture session.
2. Rust Core: Handles bounded media queues, H.264 normalization, MPEG-TS muxing, and encrypted SRT transport.
3. Data Flow: Video is captured via Meta DAT, encoded using Android's H.264 hardware encoder, processed by the Rust core, and streamed via encrypted SRT to a live ingest target provided by the developer's backend.

Who it’s for

Developers building first-person video applications for smart glasses, such as remote expert support, hands-free field work, live training, and AR/VR companion experiences.

Highlights

• Low Latency: Uses GOP-aware backpressure and bounded queues to prevent memory growth and latency spikes on weak networks.
• Resilient Connectivity: Features configurable SRT reconnects with jitter and timeouts to maintain a stable stream.
• Hardware Optimized: Leverages Android hardware H.264 encoding for efficiency.
• Developer Friendly: Provides a clean Kotlin API using Flow and StateFlow and includes an Android simulator for testing without physical hardware.