hyperaudio / ha-converter

Hyperaudio Converter - converts from JSON/SRT to HTML Based Interactive Transcript
https://hyperaud.io/converter/converter.html
12 stars 12 forks source link

🔰 Request: Support Importing Deepgram / Sonix AI JSON #22

Open natelawrence opened 3 months ago

natelawrence commented 3 months ago

In a recent hunt for more ASR providers who offer per-word timecodes I found some that I already knew of and a few I hadn't heard of before. Among all providers are Deepgram and Sonix AI.

I'm aware that HyperAudio Lite Editor already has Deepgram integration and converts Deepgram JSON back to HyperAudio hypertranscripts, but it would be good to have that functionality explicitly added to HyperAudio Converter for people who have previously exported Deepgram JSON and would like to import it to HyperAudio Lite Editor without paying Deepgram a second time for transcription.

I'm attaching an example Deepgram JSON file below. ASR Timed Text Format Test 2 [Deepgram Nova 2].json The corresponding audio file can be obtained here.


My testing also revealed that Sonix AI currently uses Deepgram as their backend ASR technology, so being able to convert Deepgram's format should mean that Sonix AI customers should be able to convert their Sonix AI transcripts for further editing in HyperAudio Lite Editor, if needed.

Sample Sonix AI format data can be obtained here: ASR Timed Text Format Test 2 (View Link) ASR Timed Text Format Test 2 (JSON)