File Size Limitation on Video/Audio to Text

Hey fellow Gliders,

I have set-up an LMS platform for a customer where they upload video-lessons into a Glide Column. The audio from the video is then automatically transcribed to text using the Glide Audio To Text AI / OpenAI integration.

However there’s a limitation on the file size. As soon as the video is too large, the transcription service does not work. I’ve allready tried to Google Cloud integration for audio transcription but that doesn’t seem to work either.

Uploading to YouTube + using the YouTube Transcription API is too much hassle for them. Does anybody else have a solution for this? Perhaps an external API that we could connect and does it directly work with the Glide urls of the uploaded video file?

Many thanks again for your support!

Best regards,


How should I handle this

You may try using WebM files. I haven’t looked into it, but the file size is very small.

1 Like

I would try DeepGram, they’re one of the fastest out there.

1 Like

Hi ThinDinh Many Thanks!!!

I tried deepgram and it works as a charm.
You get $200 free budget to start.
File size is 80x that of whisper and it just works with mp4 (loom video’s uploaded in Glide)
I did need to add the detect language component to make it work in Dutch.
And Dutch transcription IMO seems to be lower quality than that of normal openai whisper.

1 Like

What model are you using?

So have you used their API yet or just the web interface?

Their api via the api column in glide

1 Like

I really want OpenAI to work with Groq for this reason. Their model has a great output quality, but sometimes it takes forever to get a result.

1 Like

I assume with the 4.o release OpenAI API’s should be faster as well?

That is what’s expected, but I haven’t used it much to compare with the older models.