Zoom Closed Captions

This is a simple Python script that uses speech recognition libraries to post 3rd-party closed captions for Zoom meetings.

Table of Contents

• Usage
• Development
  • Mac
  • Windows - not tested
  • Ubuntu - not tested
• Other Useful Resources

Usage

Right now only Mac OSX is supported without following the development installation instructions.

0. Download & unzip the latest release of zoom_cc.

1. Enable closed captioning in Zoom settings. Refer to Zoom Help Center for instructions.

2. Start a meeting and get the API token

3. Configure zoom_cc_v0.1_OSX/settings.json for your Zoom meeting

• zoom_api_token = Paste the copied API token. If left empty, you'll be prompted for the token at runtime
• seq_count = 0 for a new meeting, unless you are restarting the script during the same meeting you were already captioning. This increments with each caption so the Zoom API knows whether it received a new caption.
• lang = Language option in post request (e.g. 'en-US' for English, 'zh' for Chinese)
• mic_timeout = Listening timeout for recognizer before resetting because nothing was said
• phrase_time_limit = Speaking timeout before transcribing and posting to Zoom (set lower to update Zoom more often, higher for more continuous phrases)

4. Start recognition script

• Open up Terminal
• Open up Finder to where zoom_cc_0.1_OSX is downloaded. Click and drag zoom_cc_v0.1_OSX/dist/zoom_cc_v0.1_OSX to the Terminal Window
• Press enter to run

5. When prompted, type y to load settings from settings.config. Otherwise paste API token via command line with ⌘V when prompted.

6. Check Zoom for transcriptions. Words should be appearing on the bottom of the screen and in the full transcript.

7. Use ctrl+C to exit. The settings.json will be updated with the last used token and sequence count in case the script needs to be restarted during a Zoom meeting.

Development

Mac

1. Get Hombrew

2. Install Python3 & PyAudio prerequisites

brew install python portaudio

5. Get PipEnv

pip3 install pipenv

6. Clone this git repo

git clone https://github.com/heyyeh12/zoom_cc.git

7. Use PipEnv to install other dependencies

cd zoom_cc
pipenv install

8. Run same usage steps as above, except start the Python script instead of executable in step 4

Using SpeechRecognizer Google Speech Recognition

pipenv run python speech_recognizer_closed_caption.py

(Experimental) Using Google Cloud Speech API. NOTE: requires setting up GOOGLE_APPLICATION_CREDENTIALS

pipenv run python transcribe_streaming_infinite.py

9. Create executable with PyInstaller

pipenv run pyinstaller --onefile speech_recognizer_closed_captions.py

Windows - not tested

1. Install Python3

2. Clone this git repo

git clone https://github.com/heyyeh12/zoom_cc.git

3. Install PyAudio system-dependent prerequisites

python -m pip install pyaudio

4. Get PipEnv

pip3 install pipenv

5. Use PipEnv to install other dependencies

cd zoom_cc
pipenv install

For Ubuntu - not tested

1. Install Python3

2. Clone this git repo

git clone https://github.com/heyyeh12/zoom_cc.git

3. Install PyAudio system-dependent prerequisites

sudo apt-get install python-pyaudio python3-pyaudio

4. Get PipEnv

pip3 install pipenv

5. Use PipEnv to install other dependencies

cd zoom_cc
pipenv install

Other Useful Resources

• SpeechRecognition Python module
  • microphone/recognizer settings
• Google Cloud Speech-to-Text API