Getting Started

Step 1: Add the stage to your bot or chat

Statosphere needs to be added before it does anything. You have two options:

• Per-chat: Open a chat's settings, find the stage dropdown, and search for "Statosphere." Add it to that specific chat.
• Per-bot (recommended for sharing): On your bot's page, scroll down to the Stages section and add Statosphere there. This applies it for everyone who starts a new chat with your bot.

The creator recommends testing in a specific chat first before applying to a public bot.

What a settings file looks like

Statosphere is controlled by a settings file in JSON format. JSON looks like this: field names on the left in quotes, their values on the right. You do not need to write it by hand — the visual editor (below) builds it for you. What matters is knowing that the "Configuration" field in Statosphere expects this kind of text when you paste your settings in.

Step 2: Open the config editor

Statosphere does nothing without configuration. Rather than handwriting JSON, use the official external editor:

lord-raven.github.io/statosphere-editor/

Build your setup in the editor's GUI, then click "Copy" at the bottom to get the JSON.

Step 3: Paste into the stage config modal

1. In Chub, open the chat or bot settings and find the Statosphere stage.
2. Open its configuration modal. You will see a single text field labeled "Configuration."
3. Paste the JSON from the editor into that field.
4. Save.

Step 4: Refresh the chat

Statosphere reads the config when the page loads. After pasting, refresh the chat page to pick up the new config.

A minimal working example

Here is a complete config you could paste right now. It tracks a mood variable and flips it when the classifier detects positive or negative sentiment in the user's message.

{
  "variables": [
    {
      "name": "mood",
      "initialValue": "\"neutral\""
    }
  ],
  "classifiers": [
    {
      "name": "Sentiment",
      "inputTemplate": "{{content}}",
      "inputHypothesis": "This message expresses {} sentiment.",
      "classifications": [
        {
          "label": "positive",
          "category": "sentiment",
          "threshold": 0.6,
          "updates": [
            { "variable": "mood", "setTo": "\"happy\"" }
          ]
        },
        {
          "label": "negative",
          "category": "sentiment",
          "threshold": 0.6,
          "updates": [
            { "variable": "mood", "setTo": "\"unhappy\"" }
          ]
        }
      ]
    }
  ],
  "content": [
    {
      "category": "Stage Direction",
      "condition": "true",
      "modification": "\"The user's current mood is: \" + mood"
    }
  ]
}

Why the double quotes in "\"neutral\""?

Every formula field in Statosphere is a string of code, and the inner \" marks tell it "this is a literal word, not a variable name." The double-quoted "\"neutral\"" is a formula that evaluates to the string neutral. See the variables page for the full story — just copy the pattern for now.

What this does:

• Starts with mood = "neutral".
• After each user message, the classifier checks whether it sounds positive or negative.
• Whichever label wins (above threshold: 0.6 — meaning the classifier needs to be at least 60% confident) sets mood to the matching string.
• A Stage Direction injects the current mood as a hidden instruction before the bot replies.

Editing an existing config

Drop your current config JSON back into the editor to modify it, then copy and re-paste. The editor understands the full config shape.

Debugging

If something is not working, open your browser's developer console: press F12, then click the Console tab at the top of the panel that appears. Look for any red error messages — those will tell you if your config has a problem. Statosphere also logs what it loaded, what classifiers fired, and what values your variables have after each turn.

See Debug Mode for more detail on reading the console output.