This looks fantastic, is there a feature to copy the text to clipboard?

Hello everyone,

I want to use deepSpeech to type some of my emails. I’m super excited about DeepSpeech project as an alternative to the Google voice typing and voice transcription service. Been a Mozilla fan forever since before firefox 1.0 and I’m so happy Mozilla worked on DeepSpeech. Such an important project for decentralised internet and privacy.

It’s such a time saver and i love the transcription feature and am looking at having the same but using DeepSpeech :rocket:

So far I followed instructions in this video and tweaked them to come up with this list of commands on linux ubuntu 20.04:

pip install deepspeech
git clone https://github.com/mozilla/DeepSpeech-examples.git DeepSpeech-examples
git clone https://github.com/mozilla/DeepSpeech.git DeepSpeech
cp DeepSpeech-examples/mic_vad_streaming/ DeepSpeech/ -R
cd DeepSpeech
cd mic_vad_streaming/
pip install -r requirements.txt
sudo apt-get install python3-pyaudio
pip install webrtcvad
pip install halo
pip install scipy
python3 mic_vad_streaming.py -m ../../deepspeech-0.8.2-models.pbmm -s ../../deepspeech-0.8.2-models.scorer 

And after talking I can see the text in the terminal! Amazing!

Now i’d like to ideally have a feature to activate it within the browser. Do I maybe need to create a Firefox extension to get the speech to text and insert it in the TEXTAREA in the web pages I use?

Thanks for your help and any insights, maybe this already exists?

There were bugs on that to add DeepSpeech support to WebSpeech implementation, but this has been on hold for nearly a year now, and I doubt I can manage enough time to re-continue that work.

Hi @lissyx - Thanks for your answer, appreciated. I realised actually what i’m looking for, is a way to replace the “Speech to text” button in the google keyboard (which sends voice to google), and use Deepspeech instead running locally.
Ultimately i’m really seeing how Speech to text is going to be used by the mass soon. Google keyboard button for speech to text is really amazing, but having the audio sent to Google is not do-able long term.

Do you maybe know of any place where such functionality has been discussed maybe?

(I’d be willing to donate 1,500 USD to the advancement of such project if it exists.)

(i realised I asked initially in this thread about inputting the text in a textarea, which would also be awesome and i’m also interested in any other links where this was discussed, but ultimately it would be even easier for me to have a replacement for the transcribe google keyboard button. )

I fear this is way above the scope of what we do: you need to change google’s codebase there.