Skip to content


Subversion checkout URL

You can clone with
Download ZIP
This is just the core architecture of SiriServer without any Plugins or thrid party code
Latest commit dcc028c @Eichhoernchen added some OS X notes
Failed to load latest commit information.
biplist Forgot biplist license text
keys added current version using twisted http instead of asyncore, with be…
plugins Removed 'He' from register Dutch Translation - plugins/examplePlugin/…
siriApi added some spaces
siriObjects added fmfObjects
.gitignore Ignoring plugins directory added current version using twisted http instead of asyncore, with be…
LICENSE Added own license fixed error when no plugin matches, also moved to debug and using new… added some OS X notes Fixed assistant parsing for Mac OS X dictation clients, renamed phone… fixed ios5 support, started adding sync support Typo
apiKeys.conf-EXAMPLE added current version using twisted http instead of asyncore, with be… updated slight protocol changes for ios6 added some checks if flac is already disposed fixed not closing urlopen open, now using context manager to do so fixed priority reset on next run, thx to theospears (f83fdfa)
plugins.conf-EXAMPLE added current version using twisted http instead of asyncore, with be… added some checks if speex is already disposed

Siri Server Core

Do you like this?

If you like this piece of software you can help me by donating, I can afford new devices and so get deeper to the core of all of this. Make it even cooler. Or if you just want to give me a little credit for my work. But don't worry the code will remain free, you don't have to donate.

PayPal — The safer, easier way to pay online.


OS X 10.8 Update

The new dictate function of OS X Mountain Lion uses the same backing engine as Siri on iOS, therefore this server is able to work with OS X 10.8. If you want to use it, you can configure your Mac to use SiriServer. I added a minor fix to display a connected Mac correctly.

If you want to configure your Mac to use SiriServer, you can edit the:


It follows the same syntax as on iOS. Or you can download a Preference Pane for the System Preferences I wrote which allows you to set some important values. You can get it at: Just unzip it an install the SiriPrefsForOSX.prefPane by double clicking it. You should activate Speech & Dictation before using it, I haven't tested every scenario thoroughly.

iOS 6 Update

I cannot guarantee that the version here will work with anything pre iOS 6.

There were some changes in iOS 6. To use the server with iOS 6 you must disable Siri's Authentication mechanism. You can do this by adding (or via Cydia see below)

<key>Authentication Disabled</key>



The easiest way to do so is like this:

ssh into your device
cd /var/mobile/Library/Preferences
plutil -convert xml1 

now edit the contents (with your favorite editor) and add (between <dict></dict> underneath <keys>Account</keys>):

<key>Authentication Disabled</key>

To Difficult?

You can also install my tweak available via Cydia which takes care of all of this, also it allows you to select more languages. Add to your Cydia sources. And install "SiriServer Tweaks". Check your for further settings.

What is this?

This is a very early version of a Siri Server (not a proxy).

Apple's Siri is an voice controlled assistant on iPhone 4S.

With jailbreaking you can install it on other iDevices. However, Siri needs a server to communicate to do the speech processing. Apple only allows 4S devices on their servers.

This project tries to recreate the Apple Siri Server to use it with other iDevices.

You don't need any 4S keys to make it work, as it is independent from Apple.

It uses Google Speech-To-Text API. And therefore we are currently limited to commands that are shorter than 10 seconds (maybe we can overcome this).

Setup, Notes and Instructions

Install audio libraries

For the audio handling you need libspeex and libflac

On Linux simply install it via you packet manager e.g. (or see instructions and note for OS X):

sudo apt-get install libspeex1 libflac8

On OS X download libspeex and libflac from the websites above (the sources, not the binaries) and compile and install them, or simply follow the following steps:

tar -xf speex-1.2rc1.tar.gz
cd speex-1.2rc1
sudo make install
cd ..

wget -O flac-1.2.1.tar.gz
tar -xf flac-1.2.1.tar.gz
cd flac-1.2.1
./configure --disable-asm-optimizations
sudo make install

Note: you can also install libspeex via MacPorts, but libflac is not available in 64bit through MacPorts, to make it build with 64bit support you need to supply --disable-asm-optimizations in configure of libflac to make it compile

Python requirements

As this project is coded with python you need a python interpreter (this is usually already installed). I work with python 2.6.6 and 2.7.2 and both work.

You also need some python packages to make it work:


pyOpenSSL is also a requirement for twisted, so installing twisted will already force an installation of pyOpenSSL.

On a debian based system twisted can be installed via apt:

sudo apt-get install python-twisted

On OS X you can install it via easy_install (sudo easy_install pyOpenSSL twisted) or via MacPorts (sudo port install py27-openssl py27-twisted)

Certificate Generation

You also need to generate certificates for this server, they must be placed in the keys/ directory, there are dummy files to show you the correct names.

Installing API Keys

To allow plugins to reuse API keys, there is an apiKeys.conf in the root directory of the server. (Note in this git there is only a dummy file -EXAMPLE)

The general format is as follows:


The apiName is usually printed in error messages when you miss a certain API Key.

Running the server

Now you are ready to go, start the server with:


You don't need to run it as root, as we use https port 4443. If you want to use another port use:

python --port [PORTNUM]

Note: for ports <= 1024 you need to run the server as root (e.g. via sudo)

Common Errors

If we had the mid 90s this section would glow and sparkle to get your attention. There are some errors that might occur even though you did everything that was written above...

The server just crashes after a SpeechPacket

You are running Linux right? Probably debian? There is probably already a libspeex on your machine which is optimized for SSE2 which does not work with python (reason???) Check if there is a /usr/lib/sse2/

Option A: delete it (there should also be a version in /usr/lib if you installed via apt, or in /usr/local/lib if you compiled by hand)

Option B: ToDo

I cannot get a connection from device to server

Do you access your server over the internet? You need to setup your firewall and NAT to allow traffic for tcp port 4443 directed to your server Do you have a local firewall on the machine running the server? Also check if tcp port 4443 is allowed for incomming connections You must also make sure to setup the corret server and port in the spire configuration:


There is something with SSL in the error

Have you installed the ca.pem file on your phone? Do you have more than one CA certificate installed for the same domain?

=> Try deleting all certificates on the device and install the one created by gen_certs

Can I somehow verify the correct certificate? YES!

start, then take your ca.pem you think belongs to your servers certificate and run:

 echo | openssl s_client -connect [DOMAIN]:4443 2>&1 | sed -ne '/-BEGIN CERTIFICATE-/,/-END CERTIFICATE-/p' | openssl verify -CAfile keys/ca.pem 

Make sure to replace [DOMAIN] with the actual domain of the machine running (e.g. an IP address) If your ca.pem matches your server certificate you should see stdin: OK as output!

OK, what else? We can also setup a small test server using openssl to check if SSL is working (and to check if the iPhone correctly validates the server certificate):

openssl s_server -cert keys/server.crt -key keys/server.key -accept 4443 -state

When you run this (SiriServer should NOT run) it opens a basis server on port 4443 using your servers certificate.

Now you can connect with your iPhone as if you would use Siri (of course Siri won't work, we are just testing the SSL layer) It should output something like this, note the Ace http request near the end:

 Using default temp DH parameters
 Using default temp ECDH parameters  
 SSL_accept:before/accept initialization
 SSL_accept:SSLv3 read client hello A
 SSL_accept:SSLv3 write server hello A
 SSL_accept:SSLv3 write certificate A
 SSL_accept:SSLv3 write server done A
 SSL_accept:SSLv3 flush data
 SSL_accept:SSLv3 read client key exchange A
 SSL_accept:SSLv3 read finished A
 SSL_accept:SSLv3 write change cipher spec A
 SSL_accept:SSLv3 write finished A
 SSL_accept:SSLv3 flush data
 Secure Renegotiation IS supported
 ACE /ace HTTP/1.0
 User-Agent: Assistant(iPhone/iPhone3,1; iPhone OS/5.0.1/9A405) Ace/1.0
 Content-Length: 2000000000


If you followed every step of the installation and you still need help to get SiriServer up and running, join #SiriServer channel on Freenode (IRC).


A big thanks to Applidium and also plamoni for his SiriProxy which inspired me Thanks to everyone that contributed code or ideas.


This is free software. You can reuse it under the terms of the Creative Commons Attribution-NonCommercial-ShareAlike 3.0 license. So you can do what ever you want with it. But you are not allowed to sell it. Or use it commercially to make profit. If you like to do more than the license allows (e.g. run a commercial server and charge people for the use of it), please contact me and ask for a special commercial license.


Apple owns all the rights on Siri. I do not give any warranties or guaranteed support for this software. Use it as it is.

Something went wrong with that request. Please try again.