Speech Recognition

​

Speech Recognition Input Plugin

The Speech Recognition Input Plugin in OM1 provides the ability to recognize and transcribe spoken language. This plugin captures audio input from a microphone and converts it into text, making it available to the agent’s runtime core for decision-making.

​

ASR Components

The ASR components are responsible for capturing audio input from a microphone and converting it into text. The main components are (take GoogleASRInput as an example):

google asr input plugin code

​

Class Diagram

In order to simplify the diagram, we only show the most important classes and their relationships.

​

Data Flow

Speech Recognition Input Plugin Data Flow:

​

Example configuration

"agent_inputs": [
    {
      "type": "GoogleASRInput",
      "config": {
        "language": "en-US"
      }
    }
  ]

​

configuration parameters

Here is an example of how to configure the ASR input plugin:

# Example usage
config = SensorConfig(
    api_key="your_api_key",
    language="en-US",
    rate=48000,
    chunk=12144,
    microphone_name="your_mic"
)
asr_input = GoogleASRInput(config)