Defines the standard plugin API (janus_plugin struct with init, destroy, handle_message, incoming_rtp, etc.) and the core callbacks (janus_callbacks). This interface allows any third-party developer to create new plugins that integrate seamlessly with the Janus core. It is self-contained, versioned, and reusable for building custom WebRTC media processing modules.

Provides a comprehensive SDP parser, generator, and manipulator. Functions like janus_sdp_parse, janus_sdp_generate_offer, and janus_sdp_generate_answer_mline are used across all plugins for WebRTC offer/answer handling. The module is isolated (depends only on glib and jansson) and can be extracted for any project needing SDP processing.

Implements a generic media recorder that saves RTP frames into a custom .mjr container format. The API (janus_recorder_create, janus_recorder_save_frame, janus_recorder_close) is used by multiple plugins for recording audio/video streams. It is self-contained and could be reused in other media applications needing a simple recording format.

Provides a lightweight reference counting mechanism (janus_refcount) with atomic operations for thread-safe lifecycle management. Used extensively across the codebase for sessions, SDP objects, and transport handles. This is a small, reusable utility that can be dropped into any C project needing refcounting.

An audio level limiter adapted from the WebRTC project, with SIMD-optimized implementations (AVX2, SSE4.2) and runtime CPU detection. It provides a clean API (WebRtcAgc_Process, etc.) for limiting audio signal peaks. This component is standalone and can be reused in any audio processing pipeline requiring peak limiting.