Skip to content

Twitch service


This tutorial is out of date. Please check the tutorials overview for our latest tutorials.

In the UI as well as using the chat interface to send messages, you can also select Twitch channels and perform sentiment analysis on the messages published there:

Twitch channels

💡 Key ideas

The key ideas on this page:

  • A Quix service can use external APIs to retrieve data and then publish that data into a Quix topic
  • Publish time series data

Twitch credentials

To run the Twitch service you'll need to provide your Twitch API credentials. You can configure these as secret variables. The credentials required are shown in the following screenshot:

Twitch credentials

What it does

The Twitch service uses the Twitch API to read messages from some of the most popular channels. It then publishes these messages to the output topic, messages.

In the following code you can see that a time series object is created, with a timestamp, the chat message, and other data, and then published to the output stream:

def publish_chat_message(user: str, message: str, channel: str, timestamp: datetime, role: str = "Customer"):
    timeseries_data = qx.TimeseriesData()
    timeseries_data \
        .add_timestamp(timestamp) \
        .add_value("chat-message", message) \
        .add_tags({"room": "channel", "name": user, "role": role})

    stream_producer = topic_producer.get_or_create_stream(channel)

The message format on the output messages topic:

  "Epoch": 0,
  "Timestamps": [
  "NumericValues": {},
  "StringValues": {
    "chat-message": [
      "@CaalvaVoladora Boomerdemons is also up"
  "BinaryValues": {},
  "TagValues": {
    "room": [
    "name": [
    "role": [

🏃‍♀️ Next step

Part 6 - Customize the UI