A digital audio workstation written in Rust, laid out as a cargo workspace so the audio model stays clean of any backend or UI.

The crates

• rdaw-core: the pure DSP data model. audio buffers, the node trait, the process graph, transport, automation, timeline, tempo, and a handful of built-in nodes. no audio-backend dependency at all, so everything here can be driven from a test or an offline renderer
• rdaw-engine: the real-time host that drives the core graph against a live audio device
• rdaw-io: file i/o, including wav roundtrip and an offline bounce example
• rdaw-gui: an eframe/egui frontend with a timeline and transport, plus rfd for native file dialogs
• rdaw-app: the top-level binary that wires it together

The split that matters

the core is sample-format-agnostic on the inside, planar f32 everywhere, and the host converts to whatever the device wants at the very edge. keeping the DSP model free of cpal means the whole audio engine is testable offline. the test suite reflects that: automation, fades, filters, mute/solo, resampling, tempo, and timeline all have their own integration tests that run with no sound card in sight.

rdaw-core  (pure dsp, no backend)
   ├── rdaw-engine  (real-time host)
   ├── rdaw-io      (wav, bounce)
   └── rdaw-gui     (egui timeline + transport)
          └── rdaw-app  (the binary)

Run it

cargo run -p rdaw-app