Eichhoernchen is working on an open-source Siri server. Not a SiriProxy server, but a full Siri server that replaces Apple’s own Siri server. The server is currently in a very early version and is not yet ready to be used. But if the server gets updated, it might someday make SiriProxies obsolete.
By now, you probably all know that you can get Siri on your iPhone 4, iPod Touch and even your iPad with Spire. The problem is that you need a SiriProxy with iPhone 4S keys to use Siri. The reason behind this is that Siri streams your voice to Apple’s servers everytime you say something. It then processes your voice and returns text. But without keys from an iPhone 4S, Apple’s server won’t process your request.
What the open-source Siri Server does is simple: It doesn’t send your voice to Apple, but to Google’s Speech-To-Text API. That way you don’t have to worry about authentication, because you are not communicating with Apple’s servers at all!
While this might sound very simple, there are a couple of problems the developers of SiriServer are currently working on. The first one is that the Google Speech-To-Text API only allows commands shorter that 10 seconds. If you would simply cut the stream in pieces every 10 seconds, you might get false results (Instead of “What’s the weather like?” it would be “What’s the we” “ather like?” and Google couldn’t convert it correctly.) One possible solution for this problem would be to analyze the audio and trying to find short silence periods. The Audio could then be split into two parts without cutting off words.
Another problem is the processing of the commands. If you ask Siri “What’s the weather like?”, the server will match your sentence and know what to do: Send you the forecast for the next few days. It then has to get your location, request the forecast from another server and then send it back to your phone. The open-source SiriServer has to learn all of these commands. This means the developers have to program every command Siri understands. And this takes a lot of time.
We will probably have to wait a little longer until we see a working open-source SiriServer, but this project looks very promising.
You can find the source code of Eichhoernchen’s SiriServer on Github.
I already have this running. Customizing the commands is done thru python, so its pretty simple for someone with no so much knowledge of code. It is not for the faint of heart tho.
Regards!
im having trouble generating the cert. can u help me? i keep getting the response
bash: cd: gen_certs: No such file or directory
Hey Brian you have to go to his github and download the files and place all the files in your homefolder then try
Been running this for a few days now, its very easy to add to siri in your own way.
Can someone do write or record a tutorial for this? I’m not familiar with coding and using terminal. Instructions for the Mac would be great.
i have server up how do u install plugins and how do u use the api keys
I got a server setup yesterday,only so I can add personal commands 2 Siri.
Does this server work with a non-jailbroken iPhone? Thanks in advance.
This works brilliantly on my Iphone 4. The big differences are it is much faster, the downside is the replies from Siri are a bit rude, I don’t get this as I have no Siri plugins installed lol. I asked Siri ” What is the 24 hour weather forecast for #my area#?” to which the reply was “It’s raining, what is it to you?” I then said “Please don’t be so rude,” and Siri replied “Sorry, I will try to be nice.”
Certainly more interesting than the standard Apple replies lol.
…and yes those commands were too long but we were playing to see the actual capabilities
What I really would love to see is an assistant which learns. Is that too War Games? Why can’t I tie my own spoken phrase to a command in the OS; i.e. teach Siri what I meant by [something she didn't understand.] This seems like it shouldn’t be much more complicated than programming macros.
A man can dream …