What is cloud-enhanced transcription?

XRAI Glass® utilizes cloud-based providers to power its transcription and translation services.

Our current cloud-enhanced transcription partners are:

Microsoft Azure

• Language Offerings: Azure provides extensive multi-language support in over 140 languages and dialects.

• WER (Word Error Rate): Azure's cloud-based model leads to a lower WER, generally around 5 words per 100, offering high accuracy.

Amazon Web Services (AWS) (Cloud-based transcription and translation)

• Language Offerings: AWS offers wide-ranging language support compatible with XRAI's offering.

• WER (Word Error Rate): AWS's cloud-based engines also maintain a low WER, close to 5 words per 100.

• Speaker Identification Available: See how to set it up here.

Deepgram 

• Language Offerings: Deepgram specializes in transcription and supports various languages through XRAI..

• WER (Word Error Rate): Deepgram's cloud-based engines also maintain a low WER, close to 5 words per 100.

• Speaker Identification Available: See how to set it up here.

• Use Case: Best for users focusing solely on transcription without the need for translation.

VOSK (offline model)

available on Android only for now

• Language Offerings: VOSK's offline capability might limit its language offerings compared to the cloud-based engines.
• WER: VOSK's offline model results in a higher WER, with more than 20 words per 100, leading to potential dissatisfaction.
• Use Case: Primarily recommended as a backup for emergency situations where no connectivity is available. Using VOSK as the primary engine may result in decreased transcription accuracy depending on the audio quality and specific use case.