.NET dev struggles with ONNX for speech-to-text: Is it too complex?

SecondCobra · reddit · 2026-08-25

A .NET developer implementing a speech-to-text app finds the ONNX route overly complex compared to Python. Simply downloading models like Whisper or Wav2Vec2 isn't enough; extensive pre-processing code is required. The user asks if they are missing a library that handles pre-processing or a model with better native .NET support.

Original post →

More from Infra

Infra channel →