Speech Recognition Service

A very simple interface for speech recognition. This first version just uses Wit.ai as a backend.

Requirements

Docker environment running

Instructions

Use ./service to build and run a container with the app
- Create a file named local.json with the values for the Wit.ai access token and docker registry to use, such as:
```
{
  "witAccessToken" : "XXXXXXXXXXXXXXXXXXXXXXXXXXXXX",
  "dockerRegistry" : "myuser/speech"
}
```

Tips

Record audio using the console

sox -d -b 16 -c 1 -r 16k amy_short.wav just press CTRL-C when done

HTTP Requests

Look at httpie as a replacement of curl to maintain your sanity.

API

Speech

> http "http://$(docker-machine ip dev):3000/v1/speech" < data/amy_short.wav

HTTP/1.1 200 OK
Connection: keep-alive
Content-Length: 227
Content-Type: application/json; charset=utf-8
Date: Sun, 24 Jan 2016 11:06:09 GMT
ETag: W/"e3-ex14DLM3GovwJx0RhA39vw"
X-Powered-By: Express

{
    "_text": "where is amy",
    "msg_id": "ac7e9608-bad1-4ffa-a146-742595e82514",
    "outcomes": [
        {
            "_text": "where is amy",
            "confidence": 0.551,
            "entities": {
                "contact": [
                    {
                        "suggested": true,
                        "type": "value",
                        "value": "amy"
                    }
                ]
            },
            "intent": "locate_people"
        }
    ]
}

Text

> http "http://$(docker-machine ip dev):3000/v1/text?q=Where is Jennifer?"
HTTP/1.1 200 OK
Connection: keep-alive
Content-Length: 244
Content-Type: application/json; charset=utf-8
Date: Sun, 24 Jan 2016 11:09:17 GMT
ETag: W/"f4-esag/jfwHvN63/YypjT85g"
X-Powered-By: Express

{
    "_text": "Where is Jennifer?",
    "msg_id": "fb8053fd-6897-44c0-ba97-be62cf10df3b",
    "outcomes": [
        {
            "_text": "Where is Jennifer?",
            "confidence": 0.551,
            "entities": {
                "contact": [
                    {
                        "suggested": true,
                        "type": "value",
                        "value": "Jennifer"
                    }
                ]
            },
            "intent": "locate_people"
        }
    ]
}

jsantiagoh/speech-api