Monday, October 29, 2018

The State of Text-to-Speech Technology

If you've been enjoying the many text-to-speech voices available inside our Read Aloud browser extension, you may have gotten the false impression that these voices are free.  They're not.  In fact, they're quite expensive.  In the table below, we explain how we're able to provide some of these voices for free, and why we're charging for others.

Voices How much it costs How we provide it
Google voices These voices are built into the Chrome browser and are free. You can use them as much as you want.
Google Translate voices These voices are not free. In fact, Google never intended to make them available to 3rd party applications. But thanks to open-source software, we're able to bring them to you. Google might decide to cut off access to these voices in the future, however.
Google Standard and Google Wavenet voices These voices are not free. Standard voices cost $4 per million characters, Wavenet voices cost $16 per million characters.

(Google has a generous free tier if you have a Google Cloud Platform account.  Read more below.)
We're able to bring you these voices for free courtesy of Google's promotion. We don't know when this promotion will end. So these voices could go away at any time.
IBM Watson voices These voices are not free. Watson voices cost $20 per million characters. We're able to bring you these voices for free courtesy of IBM's promotion. We also don't know when this promotion will end. So these voices could go away at any time.
Amazon Polly voices These voices are not free. Polly voices cost $4 per million characters.

(Amazon has a generous free tier if you have a Amazon Web Services account.  Read more below.)
There is no promotion for Amazon voices. But thanks to open-source software, we're able to serve them to you from our cloud servers at a significant discount. You can access them via the Premium Voices in-app purchase.
Microsoft Desktop voices These voices are free if you have a Windows computer. Follow these instructions. We're serving Microsoft Desktop voices to you from our cloud servers. And to cover operational costs, we have to put them behind the paywall. You can access them via the Premium Voices in-app purchase.


As you can tell, these prices are not even close to being affordable. And you can tell that we're doing our best to bring these voices to you free of charge. Only for Amazon and Microsoft voices, we have to put them behind a paywall to cover our operational costs.

Custom Voices


We have implemented a "Custom Voices" feature, allowing you to use your own Google Cloud Platform (GCP) or Amazon Web Services (AWS) account to pay for your own usage of the Google Wavenet and Amazon Polly voices.  Most likely you will get them for free because of the generous free tiers offered by Google and Amazon.  However, the setup process is fairly involved and requires a credit card.  If there is enough demand, I will make some instructional videos to help you through it.  Leave a comment below.

If you're a school or organization, we also recommend using your own GCP/AWS account so that you have total control over the availability and cost of your usage.  With the generous free tiers, it's unlikely you'll actually incur charges.  (Note, however, that the AWS free tier expires after 12 months)