Universal-3 Pro by AssemblyAI — Speech-to-text that finally understands context
Speech-to-text that finally understands context
Promoted
Is this even related to the editor with the same name that analyses text and rates it for readability? If not, it's probably breaching various trademarks.
Report
@frassmith I worked on the product when I was at Sharethrough, which is a native adtech startup. No relation to the tool you're mentioning!
Report
For my next medium post... :)
"Shocking reasons people get fat even after exercising says Brad Pitt CEO of McDonalds" - 100
It seems like CoSchedule's Headline Analyzer goes a bit more in depth as far as metrics are concerned. What sets Hemingway apart?
Report
"Shocking results about North West that leave Kayne West and Kim Kardashian frightened from the feet up... Open your eyes because they're looking North East!" - 100 lolwut
@dylanopet Not sure about you, but I'd totally click on that.
Report
"Hey so this is something that can be interesting" 66.
Suggestions:
Increase headline length
Where's the brand?
Use more Alert Words
Talk about the body
Use Context Words
Limit the passive language
Try adding a celebrity
The suggestions seem like canned responses. But I didn't expect anything more.
Seems a little gimmicky to me. Interesting nonetheless. It's always really hard to build these kinds of NLP tools. There really no right answer. I know Buzzfeed is famous for testing the bejsus out of their headlines since they rely so much on traffic but for a hobby writer like me, it's more important for the headline to reflect the actual post. A great way to measure this (stolen from @ev and Medium) is Total Time Reading.
Back to the tool - it's a great way to market the core product (much like Hubspot did with the Website Analyzer and Headline Generator) but I don't think it delivers any true value to me.
EDIT: Didn't realize tagging someone here tags them in the tweet. Pretty neat and nicely done :)
Report
In a very objective sense, this is a quite useful tool. I've tried that several times and the results seem credible to me. One thing I would like to know is what the guys from Sharethrough mean by EEG. How do they process such data? Plus, wouldn't be nice if engagement score could also be affected by emojis? :)
Webydo
Music for Makers
Summer Bod 2020