LLMs.txt: Complete documentation index for AI agents
Make it conversational

Make it conversational

The Liveboard is up. Now for the second half of the ask: hosts can ask follow-up questions in their own words instead of hunting through Liveboard filters for "why."

Find the file in which you want to embed Spotter and import the SpotterEmbed component:

import { SpotterEmbed } from '@thoughtspot/visual-embed-sdk/react';

Step 2: Render SpotteršŸ”—

First, decide which data model hosts will query. Then:

  1. In the ThoughtSpot UI, go to Develop > Visual Embed SDK Playground.

  2. Select Spotter Embed, then select the data model you want.

  3. The Code snippet section updates with the generated code. Copy it directly into your file, or extract just the Model ID.

function renderSpotter() {
    const spotterEmbed = new SpotterEmbed(container, {
        frameParams: { width: '100%', height: '100%' },
        worksheetId: '<your-model-guid>', // or 'auto_mode', see the following note
    });

    spotterEmbed.on(EmbedEvent.Init, () => console.log('Spotter initializing…'));
    spotterEmbed.on(EmbedEvent.Load, () => console.log('Spotter loaded'));

    spotterEmbed.render();
}
Note

If your instance has Auto mode enabled, you don’t need to lock Spotter to a single model. Set worksheetId: 'auto_mode' and let Spotter pick the right data source per question.

Ā© 2026 ThoughtSpot Inc. All Rights Reserved.