Accessible Communications for Everyone (ACE) Direct is a Direct Video Calling (DVC) platform that enables direct calling from deaf and hard-of-hearing individuals to an American Sign Language (ASL)-trained agent in an organization's call center. The agent answers the call using a web browser which provides a real-time video connection to a consumer.

One of the features of the ACE Direct is the ability to display captions on both the agent and consumer portals. ACE Direct uses the IBM Watson automated speech recognition (ASR) engine. Audio from the video call is captured and sent to the IBM Watson ASR and captioned text is returned for display on both the agent and consumer portals.

Probably the best way to install the entire ACE Direct system is to start with the acedirect-public repo. Follow the documentation there for a clean install. The CHECKLISTS.md file provides an overview of the complete installation and configuration process.

The ACE Quill service provides caption support in ACE Direct. Because the ACE Quill service resides on a different server, it isn't part of the automated installation script.

Note, captions are optional in ACE Direct and will require an IBM Account and credit card for billing. Pricing information for the Watson captioning service can be found here.

1. Clone this repository onto the same server that is running Asterisk
2. Clone the dat repo in the same folder and follow the configuration instructions.
3. Download and install Node.js
4. In an elevated command prompt, run npm install -g pm2

1. The ACE Quill service uses the Google speech to text engine to support captions and requires an Google account and credit card to support billing
2. Create a Google account and create a speech to text service
3. Download a private key as a JSON from the Service Accounts menu
4. Copy the JSON file to config/google.json

ACE Quill can use IBM Watson inplace of Google for Speech to Text and Translation. If you wish to use IBM Watson instead of Google you will need to change the following files:

service.js: uncomment line 3 and comment out line 4.
api/models/acequillModel.js: uncomment line 1 and comment out line 2.

1. The ACE Quill service uses the IBM Watson speech to text engine to support captions and requires an IBM Cloud account and credit card to support billing
2. Create an IBM Watson account and create a speech to text resource
3. Download the credentials file which should have the following format:

SPEECH_TO_TEXT_IAM_APIKEY=<API KEY HERE>
SPEECH_TO_TEXT_URL=<WATSON URL HERE>

1. Copy the API key and URL into config/watson.js

1. To start the ACE Quill node server with pm2, run pm2 start process.json
2. To verify the service is running, type pm2 status and you should see output similar to this:

┌─────┬─────────────────────┬─────────────┬─────────┬─────────┬──────────┬────────┬──────┬───────────┬──────────┬──────────┬──────────┬──────────┐
│ id  │ name                │ namespace   │ version │ mode    │ pid      │ uptime │ ↺    │ status    │ cpu      │ mem      │ user     │ watching │
├─────┼─────────────────────┼─────────────┼─────────┼─────────┼──────────┼────────┼──────┼───────────┼──────────┼──────────┼──────────┼──────────┤
│ 1   │ ACEQuill-API        │ default     │ 1.0.0   │ fork    │ 4755     │ 0s     │ 0    │ online    │ 0%       │ 15.0mb   │ centos │ disabled │
│ 0   │ ACEQuill-Service    │ default     │ 1.0.0   │ fork    │ 4753     │ 0s     │ 0    │ online    │ 0%       │ 16.0mb   │ centos │ disabled │
└─────┴─────────────────────┴─────────────┴─────────┴─────────┴──────────┴────────┴──────┴───────────┴──────────┴──────────┴──────────┴──────────┘