<p>Do any libraries exist for offline voice to text in Go? Even if the accuracy isn't fantastic, I'm trying to find something that can help by always listening for a specific hot word, instead of having to press a key or a button in order to start processing voice audio. Thanks!</p>
<hr/>**评论：**<br/><br/>xlab_is: <pre><p>Any bindings to C version of CMU Sphinx would be sufficient.</p>
<p>There is an lightweight engine called pocketsphinx <a href="https://github.com/cmusphinx/pocketsphinx/tree/master/src" rel="nofollow">https://github.com/cmusphinx/pocketsphinx/tree/master/src</a></p>
<p>and this is a binding for Go:
<a href="https://github.com/shunsukeaihara/go-pocketsphinx" rel="nofollow">https://github.com/shunsukeaihara/go-pocketsphinx</a></p>
<p>I'll try to do my own bindings in a few days, the idea is cool and needs proper approach, I'm not sure how stable his version is.</p></pre>xlab_is: <pre><p>P.S. About Sphinx <a href="http://cmusphinx.sourceforge.net/wiki/about" rel="nofollow">http://cmusphinx.sourceforge.net/wiki/about</a></p></pre>DustinHeroin: <pre><p>I've taken a look at go-pocketsphinx but was not able to find a way to do a stream rather then run and exit.</p></pre>xlab_is: <pre><p>Hold on, I've almost completed my own full-featured bindings, <a href="http://github.com/xlab/pocketsphinx-go" rel="nofollow">http://github.com/xlab/pocketsphinx-go</a> with almost entire API exposed.</p>
<p>Need to write README and an example.</p></pre>DustinHeroin: <pre><p>This looks fantastic so far, well done! Looking forward to the docs. </p></pre>tclineks: <pre><p>Might try to combine the effort here <a href="https://github.com/pannous/tensorflow-speech-recognition" rel="nofollow">https://github.com/pannous/tensorflow-speech-recognition</a> with the go bindings for tensorflow (I helped hack on some but I think they're hanging out on a branch that might not get merged).</p></pre>DustinHeroin: <pre><p>That could be interesting, thanks!</p></pre>tehbmar: <pre><p>Any chance you're working on the raspberry pi version of Alexa? I had the same idea to wait on a custom word then to the listen action.</p></pre>DustinHeroin: <pre><p>That's the idea!</p></pre>chriscrossproduct: <pre><p>I had been thinking about outsourcing that part to snowboy (<a href="https://github.com/Kitt-AI/snowboy" rel="nofollow">https://github.com/Kitt-AI/snowboy</a> )....but was just in the "looking around" stage.</p>
<p>Did you find good Go library for the alexa voice services API? The API itself seems fairly straight forward to build, but if someone's done the work or started, I would rather contribute to a standing project than build something from scratch...</p>
<p>edit: fixed link</p></pre>
