May 30, 2017 | Written by: Bhavik Shah
Share this post:
IBM Watson Speech to Text API – Pricing Updates
IBM Watson Speech to Text Service has been Generally Available since July 2015 and since launching, we have received great feedback that we used to improve our service. Two major pieces of feedback that we consistently received were that “The Narrowband model usage is too costly” and “If I am using the service for convert huge volume of speech to text, can I get a discount”? We heard this feedback and are excited to announce two major changes to the pricing of the service that will help reduce the cost for our developers.
#1: Remove Add-on price for Speech to Text Narrowband Model
Until now, if you were to use the Narrowband model, the price for the add-on was $.02/per minute of audio transmitted. Along with the base price of $.02/per minute of audio transmitted, the total per minute cost to you was $.04 / per minute of audio transmitted. Starting July 2017, we will be removing this add-on charge!
So the total charge for using the Narrowband model will be $.02 / per minute of audio transmitted.
Also, like always, IBM Watson does not round the usage for every API call. So if you submit two 30 seconds of audio your total charge would be $.02.
#2: Provide Graduated Tiered Pricing
We also want to pass-along additional discounts to high-volume customers. So starting July 2017, we will be adding a graduated tiered pricing for using Speech to Text API. The way it works is shown in below:
- Usage: < 250K minutes a month: $.02 per audio minutes transmitted
- Any usage > 250K to 500K minutes a month: $.015 per audio minutes transmitted
- Any usage > 501K to 1MM minutes a month: $.0125 per audio minutes transmitted
- Any usage > 1MM minutes a month: $.01 per audio minutes transmitted
#3: Language Model Customization Feature – Generally Available
Starting July 2017, language Model customization feature will be moved from Beta to Generally Available. Once the feature is generally available the pricing for the service will be as follows:
- Service will have an add-on price of $.03 / per minute of audio transmitted.
- There will be no charge for creating or hosting custom models.
More documentation will be available on our main documentation page once the pricing change launch in July 2017.