When using vox -b voxtream "text" for the first time, the model download (~500MB) happens silently — no progress indication.
The voxtream backend in src/backend/voxtream.rs should print a message to stderr before calling the voxtream CLI:
eprintln!("Loading VoXtream2 model (first run may download ~500MB)...");
Similar to how qwen-native does it in src/backend/qwen_native.rs:35.
Small fix, one line.
When using
vox -b voxtream "text"for the first time, the model download (~500MB) happens silently — no progress indication.The voxtream backend in
src/backend/voxtream.rsshould print a message to stderr before calling the voxtream CLI:Similar to how
qwen-nativedoes it insrc/backend/qwen_native.rs:35.Small fix, one line.