A collection of command-line tools for offline speech/intent recognition

voice2json

voice2json is a collection of command-line tools for offline speech/intent recognition on Linux. It is free, open source (MIT), and supports 18 human languages.

From the command-line:

$ voice2json -p en transcribe-wav \
      < turn-on-the-light.wav | \
      voice2json -p en recognize-intent | \
      jq .

produces a JSON event like:

{
    "text": "turn on the light",
    "intent": {
        "name": "LightState"
    },
    "slots": {
        "state": "on"
    }
}

when trained with this template:

[LightState]
states = (on | off)
turn (<states>){state} [the] light

voice2json is optimized for:

• Sets of voice commands that are described well by a grammar
• Commands with uncommon words or pronunciations
• Commands or intents that can vary at runtime

It can be used to:

• Add voice commands to existing applications or Unix-style workflows
• Provide basic voice assistant functionality completely offline on modest hardware
• Bootstrap more sophisticated speech/intent recognition systems

Supported speech to text systems include:

• CMU's pocketsphinx
• Dan Povey's Kaldi
• Mozilla's DeepSpeech 0.9
• Kyoto University's Julius

Supported Languages

• Catalan (ca)
  • ca-es_pocketsphinx-cmu
• Czech (cs)
  • cs-cz_kaldi-rhasspy
• German (de)
  • de_deepspeech-aashishag
  • de_deepspeech-jaco
  • de_kaldi-zamia (default)
  • de_pocketsphinx-cmu
• Greek (el)
  • el-gr_pocketsphinx-cmu
• English (en)
  • en-in_pocketsphinx-cmu
  • en-us_deepspeech-mozilla
  • en-us_kaldi-rhasspy
  • en-us_kaldi-zamia (default)
  • en-us_pocketsphinx-cmu
• Spanish (es)
  • es_deepspeech-jaco
  • es_kaldi-rhasspy (default)
  • es-mexican_pocketsphinx-cmu
  • es_pocketsphinx-cmu
• French (fr)
  • fr_deepspeech-jaco
  • fr_kaldi-guyot (default)
  • fr_kaldi-rhasspy
  • fr_pocketsphinx-cmu
• Hindi (hi)
  • hi_pocketsphinx-cmu
• Italian (it)
  • it_deepspeech-jaco
  • it_deepspeech-mozillaitalia (default)
  • it_kaldi-rhasspy
  • it_pocketsphinx-cmu
• Korean (ko)
  • ko-kr_kaldi-montreal
• Kazakh (kz)
  • kz_pocketsphinx-cmu
• Dutch (nl)
  • nl_kaldi-cgn (default)
  • nl_kaldi-rhasspy
  • nl_pocketsphinx-cmu
• Polish (pl)
  • pl_deepspeech-jaco (default)
  • pl_julius-github
• Portuguese (pt)
  • pt-br_pocketsphinx-cmu
• Russian (ru)
  • ru_kaldi-rhasspy (default)
  • ru_pocketsphinx-cmu
• Swedish (sv)
  • sv_kaldi-montreal
  • sv_kaldi-rhasspy (default)
• Vietnamese (vi)
  • vi_kaldi-montreal
• Mandarin (zh)
  • zh-cn_pocketsphinx-cmu

Unique Features

voice2json is more than just a wrapper around open source speech to text systems!

• Training produces both a speech and intent recognizer. By describing your voice commands with voice2json's templating language, you get more than just transcriptions for free.
• Re-training is fast enough to be done at runtime (usually < 5s), even up to millions of possible voice commands. This means you can change referenced slot values or add/remove intents on the fly.
• All of the available commands are designed to work well in Unix pipelines, typically consuming/emitting plaintext or newline-delimited JSON. Audio input/output is file-based, so you can receive audio from any source.

Commands

• download-profile - Download missing files for a profile
• train-profile - Generate speech/intent artifacts
• transcribe-wav - Transcribe WAV file to text
  • Add --open for unrestricted speech to text
• transcribe-stream - Transcribe live audio stream to text
  • Add --open for unrestricted speech to text
• recognize-intent - Recognize intent from JSON or text
• wait-wake - Listen to live audio stream for wake word
• record-command - Record voice command from live audio stream
• pronounce-word - Look up or guess how a word is pronounced
• generate-examples - Generate random intents
• record-examples - Generate and record speech examples
• test-examples - Test recorded speech examples
• show-documentation - Run HTTP server locally with documentation
• print-profile - Print profile settings
• print-downloads - Print profile file download information
• print-files - Print user profile files for backup

GitHub

https://github.com/synesthesiam/voice2json