Vosk Restful Service backed by Flask and Celery

How to Setup

Clone it first

git clone git@github.com:raminious/vosk-server.git
cd vosk-server

Create Virtual Env (optional)

python -m venv env
source env/bin/activate

Download a vosk model and unzip that in the root of the project as "model" directory.
https://alphacephei.com/vosk/models

wget http://alphacephei.com/vosk/models/vosk-model-en-us-0.22.zip
unzip vosk-model-en-us-0.22.zip
mv vosk-model-en-us-0.22 model

Install requirements

pip install -r requirements.txt

Run in Development environment

python app.py
python tasks.py # in another terminal

Run in Production environment

python server.py
python tasks.py 

Endpoints

Send a new transcribe request

POST http://localhost:5000/asr/recognize
Content-Type: audio/x-wav
(wav file data)

For example using curl:

curl -XPOST -H'Content-Type: audio/x-wav' --data-binary "@files/example.wav" http://localhost:5000/asr/recognize

Response:

{ 
  id: "<task_id>" 
}

Check request status
GET http://localhost:5000/asr/recognize/status/<task_id>

Files

app.py contains the endpoints functions
task.py contains the script to run Celery and transcribing task function
server.py runs the production API service powered by gevent

Notes

This is a very basic example of using Vosk with a task scheduler like Celery. You can fork the repo and change the codes and tune Celery configs based on your requirements.