A node.js command line app to convert audio to text by VoiceBase API.
- Clone this repo and cd to the root directory.
- Install dependencies:
yarn
. - Build:
yarn build
. - Create the app config file
./data/config.json
. - Run the command:
node build/voicebase/recognize.js {audio-file} {custom-vocabulary-text-file}
.
A text file containing a comma(space or new line) separated string.
{
"voicebaseToken": "{yourVoiceBaseAPIToken}"
}
node src/voicebase/clearMedia.js
Register a SF dev account.
Create a connected app. Enable OAuth Settings. Scope: Full access.
salesforce dev account oauth password in config.js
: ${dev-password}${dev-security-token}
node build/server.js
-
if no customer found in SF, should not throw exception:
TypeError: Cannot read property 'FirstName' of undefined
-
4 SF task records were created for a single voicemail Root cause: message-sync token not saved. So it will treat the message as new again and again.
- Move to Lambda, some work is done but not finished.