deepgram/deepgram-dotnet-sdk

Supports multiple transcription modes including streaming via WebSocket, pre-recorded audio (synchronous and asynchronous with callbacks), text-to-speech, and agentic voice interactions. Built with a factory pattern for client instantiation across REST and WebSocket transports, with structured schema-based configuration for model selection and API parameters. Includes full account management APIs (projects, keys, members, usage tracking, billing) and on-premises deployment support.