anwering-machine-detection-example

This application illustrates how to use answering machine detection (amd) on outbound calls placed jambonz. To test this application configure and run it as shown below, then trigger an outbound call via REST api, e.g.

curl --location -g 'https://{jambonz_url}/api/v1/Accounts/{account_sid}/Calls' \
--header 'Authorization: Bearer {api_key}' \
--header 'Content-Type: application/json' \
--data '{
    "application_sid": "{application_sid}",
    "from": "16172223333",
    "to": {
        "type": "phone",
        "number": "16174445555"
    }
}'

Note that you need a running jambonz system along with your account_sid, api key, and the application_sid for this application, which you will have added as an Application in the jambonz portal

Configuration

After running

npm ci

You can run this application simply as

node app.js

but the following environment variables are available

name	description	default
WS_PORT	http port to listen on for incoming websocket connections from jambonz	3000
LOGLEVEL	log level (info or debug)	info
AMD_NOSPEECH_TIMEOUT_MS	time in milliseconds to wait for speech before returning amd_no_speech_detected	5000
AMD_DECISION_TIMEOUT_MS	time in milliseconds to wait before returning amd_decision_timeout	15000
AMD_TONE_TIMEOUT_MS	time in milliseconds to wait to hear a tone	20000
AMD_GREETING_COMPLETION_TIMEOUT_MS	silence in milliseconds to wait for during greeting before returning amd_machine_stopped_speaking	2000

How this application works

This is a very simple application, designed to illustrate amd and to provide code that you can use in your own applications.

An outbound call is placed to the number provided in the POST request, and when the far answers the application plays a long-ish greeting while at the same time trying to determine if it is speaking to a human or a bot. Once it determines which it is speaking to, it then switches to a different message based on whether it detected a human or a machine.

The relevant code for those interested can be found here.