Android build

For Android make sure you use mavenCentral repository and just add aar to dependencies.

repositories {

dependencies {
    implementation group: 'com.alphacephei', name: 'vosk-android', version: '0.3.32+'

In case you still want ot build android library, do the following:

cd vosk-api/android/lib
gradle build

Please note that medium blog post about 64-bit is not relevant anymore, the script builds x86, arm64 and armv7 libraries automatically without any modifications.

For an example Android application using the Vosk-API check the project.

iOS build

Available on request. Drop us an e-mail at Describe the organization you are working at if any. Describe the purpose of the project.

Python installation from Pypi

The easiest way to install vosk api is with pip. You do not have to compile anything.

We currently support the following platforms:

  • Linux on x86_64
  • Raspbian on Raspberry Pi
  • Linux on arm64
  • OSX (only x86, not M1)
  • Windows

We do not support:

  • ARMv6 (Rpi Zero is too slow)
  • OSX M1
  • Windows ARM64

Make sure you have up-to-date pip and python3 versions:

  • Python version: 3.5-3.9
  • pip version: 20.3 and newer.

Upgrade python and pip if needed. Then install vosk on Linux/Mac from pip:

pip3 install vosk

Please note that some platforms are not fully supported by pip, for example on arm64 you have install from released wheels:

On OSX M1 you can try to install x86 python in Rosetta2, vosk should work then.

If you have trouble installing, check the output of the following commands and provide it for reference:

python3 --version
pip3 --version
pip3 -v install vosk

Websocket Server and GRPC server

We also provide a websocket server and grpc server which can be used in telephony and other applications. With bigger models adapted for 8khz audio it provides more accuracy.

The server is installable from docker and can be run with a single command:

docker run -d -p 2700:2700 alphacep/kaldi-en:latest

For details see

Compilation from source

If you still want to build from scratch, you can compile Kaldi and Vosk yourself. Please note that compilation is not straightforward and includes several nuances. We recommend to follow our build process we document in Docker files:

For Windows and Raspberry Pi we recommend cross-build with mingw and corresponding ARM toolchain. See docker files for details.

The outline of the build is here.

Python module build

After you have built kaldi with openblas,clapack and openfst, you can buid python module:

git clone
cd vosk-api/src
cd ../vosk-api/python
python3 install

Usage examples


Clone the vosk-api and run the following commands:

git clone
cd vosk-api/python/example
mv vosk-model-small-en-us-0.15 model
python3 ./ test.wav

When using your own audio file make sure it has the correct format - PCM 16khz 16bit mono. Otherwise, if you have ffmpeg installed, you can use, which does the conversion for you.

Find more examples such as using a microphone, decoding with a fixed small vocabulary or speaker identification setup in the python/example subfolder

For more info see this video:


We distribute Vosk library on MavenCentral. To plug the library simply add MavenCentral repo and add the dependency:

repositories {

dependencies {
    implementation group: '', name: 'jna', version: '5.7.0'
    implementation group: 'com.alphacephei', name: 'vosk', version: '0.3.31+'

For example of a full demo project built with gradle see:

The code samples are here:

We support Java 8+ on Linux, OSX and Windows.


We recommend install with nuget. See

To try the system

git clone
cd vosk-api/csharp/demo
dotnet run

To integrate library in your software simply execute

dotnet add package Vosk

and then follow the example in our code.

For details of the wrapper see the nuget folder.

Nuget should work on Windows and Linux. We might add OSX later.


We implement node bindings with node-ffi-napi library and thus should support more or less recent node versions

To install node module simply run

npm install vosk

For example of the API see the code on github.