After all it aims to become a full fledged Siri clone. It’s written in Python, the GUI is based on PyGtk. Speech recognition is performed either using Pocketsphinx or the Google Voice Search backend.
But well, it’s rather buggy:
First off the speech recognition using Google Voice API was broken. Saera was always relying on Pocketsphinx. If that one didn’t come up with the utterance, it passed the voice recording on to Google’s API. So far, so good. But it didn’t pass the utterance back to the processing module. It just passed it on to answers.com. I’ve quickly fixed that and provided a pull request on GitHub, which was merged within two days.
I’ll definitely put some more effort into the project. Mainly I’d like to speak to Saera in German language and have a plugin based processing backend.