Speechify tests almost 1,000 AI-generated ads per day on top of roughly 8,000 human-made organic creatives per month.

“We test almost a thousand day generated ads a day right now on top of the roughly 8,000 organic creatives we make with humans every single month.”

Speechify's inference cost is single-digit dollars per million characters, versus $30 to $100 per million characters for competing models, roughly two orders of magnitude cheaper.

“We got it to the point of singledigit dollars for a million characters. If you go to any other model it'll be anywhere from $30 to $100 per million characters. There's like two order deviation standard deviations cheaper.”

Speechify: building proprietary voice inference infrastructure at single-digit dollars per million characters, well below market rates, and claiming 94% of the B2C voice agents market.

“We got it to the point of singledigit dollars for a million characters. If you go to any other model it'll be anywhere from $30 to $100 per million characters. There's like two order deviation standard deviations cheaper.”

Cliff Weitzman on Speechify token spend overtaking salaries

“Next year I expect we'll spend more in tokens than we'll spend on salaries.”

178,000 people applied to open engineering positions at Speechify in one year.

“178,000 people applied to open engineering positions at Speedify last year.”

Speechify does not assign a designer at the start of a new product build, instead using default AI-generated code and bringing designers in only after the feature is proven.

“We do not assign a designer in the beginning like we say just use the default clock code stuff and then we'll bring designers after and they'll make it a lot better.”

Speechify claims 94% of the B2C voice agents market.

“We have 94% of the BTOC voice agents market.”

60-second message response time is the single best predictor of engineering success at Speechify, outweighing conventional hiring signals.

“You said before literally 20 minutes ago 60-second response times are the best predictor for any engineer's success.”