chromedictator

https://github.com/stts-se/chromedictator

Build and run from source

Requires Go >= 1.12.

If you do not already have Go installed, download and install the most recent stable version from https://golang.org/dl/, then:

git clone https://github.com/stts-se/chromedictator
cd chromedictator
go get

To start the server:

 go run chromedictator.go

or

 go build
 ./chromedictator

Go to the following URL in Google Chrome:

 http://localhost:7654

The server will create a audio_files sub-directory in the corrent directory if it does not already exist.

The server will create a abbrevs.gob file, containing mappings from abbreviations to expanded forms, if it does not already exist.

Run from pre-built binaries

Download the latest zip file from releases, unzip, and run the binary for your OS.

Build and package pre-compiled version

The make command will generate a zip file containing everything needed to run the server, including default executables for the following operating systems:

• chromedictator (linux)
• chromedict_win
• chromedict_mac (darwin, untested)

Run pre-compiled version

1. Unzip the zip file
2. Start the server using the pre-compiled executable for your OS.
3. Start Google Chrome and visit http://localhost:7654

Record from audio output

Here's a neat trick to record from your audio output of your computer, using PulseAudio (for Linux): https://unix.stackexchange.com/questions/130774/creating-a-virtual-microphone/153528#153528

Requirements

• Google Chrome

Files ending up in the server's session folder

.webm

Audio (media) file used by Google Chrome. Can be converted into .wav or other formats using e.g. ffmpeg.

.json

Metadata file accompanying the .webm file with the following fields:

• session_id : the name of the session
• start_time : recording start timestamp (ISO format)
• end_time : recording end timestamp (ISO format)
• time_code_start : recording start time relative to session start time (milliseconds)
• time_code_end : recording end time relative to session start time (milliseconds)

Sample JSON can be found in audio_files/default/audiotst.json:

{
  "session_id": "default",
  "start_time": "2018-11-16T15:38:00.606Z",
  "end_time": "2018-11-16T15:38:03.305Z",
  "time_code_start": 12593,
  "time_code_end": 15292
}

.rec

Text file containing the original recognition result.

.edi

Text file containing manually edited recognition result. May be identical to the contents of the .rec file.