Keyword Detection
The Keyword Detection node is designed to monitor an audio stream in real time and identify whether one or more specified keywords are present. This node supports the detection of multiple keywords simultaneously and can direct the flow of an audio processing workflow based on the results.
Input
| Name | Data Type | Description |
|---|---|---|
| Audio | Audio Stream | The audio source from which the node will detect keywords. It must be connected to a valid audio input. |
| Enabled | Boolean | Activates the node. When enabled (true), the node performs keyword detection; when disabled (false), the node remains inactive. |
| Keyword Type | String | Specifies the category or type of keywords to be detected. It helps define the context for the detection process. |
| Keyword | String | Lists the keyword(s) to detect. Multiple keywords can be provided, typically as a comma-separated list, allowing the node to monitor for more than one term concurrently. |
Output
| Name | Data Type | Description |
|---|---|---|
| Detected | Exec | Executes when the node successfully detects at least one of the specified keywords within the audio stream |
| Not Detected | Exec | Executes when no specified keywords are found in the audio stream |
| Audio | Audio Stream | The original audio stream is output from the node. This output is intended for connection to another audio process node, like the Speaker node. |
| Detected Keyword(s) | String | Provides a list of all the keywords that have been detected. If multiple keywords are found, they are returned as a comma-separated list. |
How Keyword Detection Works
The node continuously processes the incoming audio stream and compares it against the specified keywords. When a keyword is detected, it can trigger subsequent actions by executing one output branch. If no keyword is detected, an alternative branch is executed. Additionally, the node can forward the original audio stream for further processing, such as playing the audio through a Speaker node.