Repository Issue Activity (beta)

#137 Request: macOS Intel (x86_64) Support

#181 Discrepancy between reproduced WER and paper results on LibriSpeech clean (Moonshine Tiny)

#148 Unable to reproduce WER results with streaming models

#169 Enhance compile compatibility

#182 Benchmark comparison request: FunASR SenseVoice (170x realtime, 50+ languages)

#125 Initial build breaks with Windows cli-transcriber example

#122 Java bindings

#129 Unexpectedly slow performance on Intel CPUs for Windows compared to AMD

#133 [Feature Request] Enable explicit GPU Execution Providers (CUDA/DirectML) on Windows

#135 Feature Request: Context Biasing / Vocabulary Prompting (similar to Whisper's initial_prompt)

#142 Add VAD API

#145 Feature request: word-level timestamps

#156 How can I use IntentRecognizer on Android?

#158 fix(python): transcript_line_t struct mismatch in PyPI v0.0.49 causes SIGSEGV on multi-line transcription

#161 Documentation on `moonshine_transcribe_without_streaming` does not match actual behaviour

#163 Can you provide an Android demo for calling the text-to-speech functionality?

#167 ARM64 int8 quantized model produces hallucinations on short utterances

#171 Proper python CLI interface

#176 Community testing: React Native Moonshine wrapper

#184 CMake fails on Linux x86_64 due to unsupported platform detection

#195 build fails on linux x86_64 with a c++ error

#196 Python: `mic_stream.add_audio()` performs synchronous transcription, blocking the PortAudio callback

#174 moonshine-voice AAR includes only arm64-v8a JNI libs (missing armeabi-v7a / x86_64 support)

#164 Can the Android SDK set minSdk = 35 to a lower value to support Android 7 devices?

#197 minSdkVersion 35 in library AAR blocks adoption on Android 14 and below