Natter

Programmable and extensible voice automation for Linux, MacOS, and Windows

MASTER	Code Quality

Linux	Windows	MacOS

It’s designed for programmers to use and extend as they see fit.

• Natter
  • Features
    • API Integrations
    • Automation Intents
    • Grammar Designs
  • Development
    • Makefile
    • Dependencies
      • Running
  • Distribution

Features

API Integrations

Use natter to turn speech into text with a number of Web Speech API's:

• Google
• IBM Watson
• Wit.ai
• Mozilla Deep Speech

Automation Intents

Automation intents are powered by the context of the application you are in. When in a .doc for example, you might want to just speak freely and note everything you say. In a .js file you will probably want access to some programming grammars...

Grammar Designs

Design grammars that can implement macros, refactor text and control window focus:

• Speech to text :: “Hi, who is there” => hi, who is there
• Speech to key bindings :: “slap down copy” => {enter_key} {down_key} {control+c}
• Transform text functions :: “sentence bring all your API’s” => Bring all your API’s
• Custom macros :: “jason key data value brackets number list 1 34 67” => { “data”: [ 1, 34. 67 ] }

Development

Makefile

The Makefile is a developers kickstart which will provide all the commands that you require.

make start
make test

Dependencies

• Google Cloud Platform Service accounts
• Sox for audio recording

# Google Cloud Platform - For streaming recognition - replace with your path :+1:
export GOOGLE_APPLICATION_CREDENTIALS="/home/simon/code/node/natter/config/credentials/natter-credentials.json"

# linux - Sox for audio recording

sudo apt-get install sox

# MacOS - Sox for audio recording
brew install sox

Running

# Install dependencies
yarn
# Run in demo mode
yarn demo
# Run in dev mode
yarn start

Distribution

Currently the binaries are built via Microsoft Azure Pipelines.

These pipelines will produce zip files which can be stored at remote URLs and downloaded.