Speechly features

Feature availability varies depending on if you are using pre-recorded audio or live streaming audio, your application language and your project payment plan.

Category	Feature	Pre-recorded	Live streaming	Availability
Models	Conformer RNN‑T models	3	3
	Whisper models	1
	Model adaptation / training	✓	✓	RNN‑T models only
	Data annotation service	✓	✓	Enterprise plans only
Transcription	Pre-recorded audio	✓	✓
	Live streaming audio		✓
	Language support	99	English
	Language detection	✓
	Model selection	✓	✓
	Word level timestamps	✓	✓	RNN‑T models only
	Punctuation	＊	＊
	Number & date formatting	＊	＊
	Silence segmentation	＊	✓	RNN‑T models only
	Interim results		✓	RNN‑T models only
	Voice activity detection	✓	✓
	Evaluate ASR accuracy	✓
Speech understanding	Intent detection		✓	RNN‑T models only
	Entity detection		✓	RNN‑T models only
	Lookups		✓	RNN‑T models only
	Text labeling	＊	＊	English language only
	Language translation	＊		Whisper models only
	Evaluate NLU accuracy		✓
Audio analysis	Language detection	✓	＊
	Audio event labeling	＊	＊
	Tone of voice	＊	＊
	Lyrics transcription	＊	＊	English language only
Supported audio formats	WAV	✓	✓
	FLAC	✓	＊
	OGG	✓	＊
	MP3	＊	＊
	AAC	?	?
	Up and down sampling	✓	✓
Deployment options	On-device		✓	Enterprise plans only
	On-premise	✓	✓	Enterprise plans only
	Cloud	✓	✓
Integration	Browser client		✓
	React client		✓
	Android client		✓
	iOS client		✓
	Unity client		✓
	Speechly decoder		✓	iOS, Android and C
	gRCP API	✓	✓
	REST API	✓
Developer tools	Dashboard
	CLI

✓ available   ＊ planned   ? under consideration