Introducing Speech 2 Text API by Google

By: Todd Fisher, 24 Mar 2011
wave

Yesterday Google announced the release of Chrome 11 beta.  As usual Google Chrome team continues to push the limits of the web.  With this release they made public their use of a little known Web API for decoding speech into text.  The new API from google is not documented and has probably been in existence since early 2009 when they opened up more of Google Voice and made know their use of transcription for voice mail.   These features are all very interesting.  Thanks to the good work from “don’t panic”, it was very easy to put together a ruby library that helps make it even easier to use the speech API.  Check out the source a github.com.

Here’s a sample of the code:

require 'speech'
audio = Speech::AudioToText.new("i-like-pickles.wav")
puts audio.to_text.inspect
# {"captured_json"=>[["I like pickles", 0.92731786] ], "confidence"=>0.92731786}

Let us know if you have some interesting ideas for how to use this.

Tags: , , , ,

This entry was posted on Thursday, March 24th, 2011 at 9:44 pm and is filed under News, Web Development. You can follow any responses to this entry through the RSS 2.0 feed. Both comments and pings are currently closed.

One Response to “Introducing Speech 2 Text API by Google”

  1. Ahmad says:
    April 5, 2011 at 1:37 am

    Hi Todd,
    That is really clever.

    I only wish it was in the form of PHP scripts, so that we could use it for Section 508 Web Compliance of our Weather Presentations (Videos that have accompanied text, based on the spoken words).