Kucedr supports dictation and spoken assistant responses through the providers and models you configure.
Dictation
Available. Streaming speech-to-text models append partial and final transcript events while recording. Batch-only models record locally and submit the audio when recording stops.
The dictation interface includes microphone permission checks, elapsed time, a waveform, confirm, cancel, and error states.
Spoken Responses
Available. Assistant messages can be read aloud using the configured text-to-speech provider.
Voice Conversation
Available. With a supported OpenAI or xAI realtime model, the empty-composer voice action starts a full-duplex session with microphone capture, assistant audio playback, transcript messages, interruption, tool activity, permissions, and generated media in the current conversation.
See Providers for supported speech models and runtime adapters.