developers.home-assistant/index.md at df7092023c2fa03505c55a9dc6d6862ac4a5ec18

* Add docs for pipelines

* Minor changes

* Fix parameter

* Fix other parameter

2023-03-20 23:59:32 -04:00

title

title
Pipelines

A pipeline runs the common steps of a voice assistant:

Pipelines are run via a WebSocket API:

{
  "type": "voice_assistant/run",
  "language": "en-US"
}

The following input fields are available:

Name	Type	Description
`intent_input`	string	Required. Input text to process.
`language`	string	Optional. Language of pipeline to run (default: configured language in HA).
`pipeline`	string	Optional. Id of a pipeline to run (default: use first one that matches specified language).
`conversation_id`	string	Optional. Unique id for conversation.
`timeout`	number	Optional. Number of seconds before pipeline times out (default: 30).

Events

As the pipeline runs, it emits events back over the WebSocket connection. The following events can be emitted:

Name	Description	Emitted	Attributes
`run-start`	Start of pipeline run	always	`pipeline` - Id of pipeline `language` - Language used for pipeline
`run-finish`	End of pipeline run	always
`intent-start`	Start of intent recognition	always	`engine` - Agent engine used `intent_input` - Input text to agent
`intent-finish`	End of intent recognition	always	`intent_output` - conversation response
`tts-start`	Start of text to speech	audio only	`tts_input` - text to speak
`tts-finish`	End of text to speech	audio only	`tts_otuput` - URL of spoken audio