VoiceContext

Copy Markdown

Voice-to-text transcription and text-to-context extraction for Elixir.

Install

def deps do
  [
    {:voice_context, path: "../lib"}
    # {:voice_context, "~> 0.1.0"}
  ]
end

Usage

{:ok, text} = VoiceContext.transcribe(audio_path_or_bytes)

{:ok, context} = VoiceContext.extract_context(text)

{:ok, %{transcript: t, context: c}} = VoiceContext.from_audio(audio_path_or_bytes)

Configuration

config :voice_context,
  speech_adapter: :local,
  context_adapter: :local,
  max_context_chars: 2000,
  default_language: "en"

Custom adapters implement transcribe/2 and extract/2.

Develop

mix deps.get
mix test
mix docs