Live Transcription for Conferencing (Closed Captions - Beta)

100ms real-time transcription engine generates a live transcript (closed captions) during a conferencing session. The SDK provides a callback with the transcript for each peer when they speak.

Minimum Requirements

• Minimum 100ms SDK version required is 1.9.0

How to implement closed captioning?

Implement on(transcripts: HMSTranscripts) from HMSUpdateListener callback like below:

    public func on(transcripts: HMSTranscripts) {
        transcripts.transcripts.forEach { transcript in
            // handle transcript
        }
    }
}

Here is an example implemenation:

    public func on(transcripts: HMSTranscripts) {
        transcripts.transcripts.forEach { transcript in
            let peerModel = transcript.peer
            
            if !(lastTranscript?.isFinal ?? false) {
                _ = self.transcriptArray.popLast()
            }
            
            if peerModel == lastTranscript?.peer {
                self.transcriptArray += [" " + transcript.transcript]
            }
            else {
                // if last transcript was not final pop the speaker label as well
                if !(lastTranscript?.isFinal ?? false) {
                    if transcriptArray.last?.contains(":") ?? false {
                        _ = self.transcriptArray.popLast()
                    }
                }
                self.transcriptArray += ["\n**\(peerModel.name.trimmingCharacters(in: .whitespacesAndNewlines)):** "]
                self.transcriptArray += ["\(transcript.transcript)"]
            }
            
            lastTranscript = transcript
        }
    }
}