This might be a "you're SOL" scenario, train your own model or pay/upload. I think there might be some self-hosted ones too, I understand someone has had to train/make it.
I'm just needing this because we're dealing with a lot of video and I don't really want to watch them. I just want to look for keywords mentioned.
Eventually we will incorporate automatic/real time translation and for that it's like we will definitely be feeding our stream to the service... so this might be a moot point but just asking for thoughts/opinions.