Skip to main content
POST
/
v2
/
project
/
{projectID}
/
user
/
{userID}
/
interact
/
stream
Interact (stream)
curl --request POST \
  --url https://general-runtime.voiceflow.com/v2/project/{projectID}/user/{userID}/interact/stream \
  --header 'Content-Type: application/json' \
  --header 'authorization: <api-key>' \
  --data '
{
  "action": {
    "type": "action",
    "payload": {
      "actions": [
        {
          "type": "<string>",
          "payload": "<unknown>"
        }
      ],
      "label": "<string>"
    },
    "diagramID": "<string>",
    "time": 123
  },
  "variables": {},
  "sessionID": "<string>"
}
'
{
  "event": "trace",
  "data": {
    "type": "audio",
    "payload": {
      "state": "start",
      "messageID": "<string>",
      "delay": 123
    },
    "paths": [
      {
        "label": "<string>",
        "event": {
          "type": "<string>",
          "payload": "<unknown>",
          "diagramID": "<string>",
          "time": 123
        }
      }
    ],
    "defaultPath": 123,
    "time": 123
  },
  "id": "<string>",
  "type": "event"
}

Authorizations

​
authorization
string
header
required

Voiceflow Dialog Manager API key (VF.DM)

Path Parameters

​
projectID
string
required

The ID of the Voiceflow project to interact with.

​
userID
string
required

An ID which uniquely identifies the user having the conversation.

Query Parameters

​
modality
enum<string>

Certain features are only available depending on your modality.

Available options:
voice,
chat,
api
​
environment
string
default:development

Which environment to target within the project.

​
audio_events
boolean

If enabled, the configured speech-to-text provider will be invoked to convert system response text to audio. Audio traces will be returned from the server.

​
audio_encoding
enum<string>
default:audio/mp3

When audio_events are enabled, selects the encoding of audio traces returned by the server.

Available options:
audio/mp3,
audio/x-mulaw,
audio/pcm
​
completion_events
boolean

When enabled, LLM traces will be broken up into streamed-in chunks - documentation

​
userTimezone
string

The timezone experienced by the user in this conversation.

​
state
boolean

When enabled, conversation state is returned in response to each interaction.

Body

application/json
​
action
object
required

Start a new conversation with a launch action or reply to a message with the text action.

​
variables
object

Set variables for the next turn of the conversation.

​
sessionID
string

Unique ID of the conversation session.

Response

200 - text/event-stream

Traces contain responses, media and metadata generated by the conversation runtime.

​
event
enum<string>
required
Available options:
trace
​
data
object
required
​
id
string
​
type
enum<string>
Available options:
event