codeforequity-at/botium-speech-processing

Provides unified REST and WebSocket APIs for speech recognition (via Kaldi) and synthesis (via MaryTTS), with built-in audio processing through SoX and support for cloud providers like Google Cloud, IBM Watson, Azure, and Deepgram via request-level credential overrides. Containerized with Docker Compose for local deployment or cloud-specific slim configurations, includes result caching, file system watchers for batch processing, and optional real-time streaming endpoints. Targets chatbot platforms like Rasa and Botium for voice-enabled IVR and automated voice app testing scenarios.

944 stars.