You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
:page-description: Learn the different ways to integrate ThoughtSpot analytics and AI experiences into your applications and AI agents, including MCP Server, Spotter Embed, and Spotter APIs.
8
+
9
+
ThoughtSpot provides several options to bring governed analytics and AI-powered experiences into your applications, AI agents, and workflows.
10
+
11
+
== AI integration options
12
+
13
+
The primary ways to integrate ThoughtSpot analytics and AI into your environment are:
14
+
15
+
* xref:ai-integration-options.adoc#_thoughtspot_mcp_server[ThoughtSpot Model Context Protocol (MCP) Server]
16
+
* xref:ai-integration-options.adoc#_embedding_spotter_in_your_app[Embedding Spotter in your app]
All options allow using your existing ThoughtSpot data models, Liveboards, Answers, row-level and column-level security, and governance. The main differences are where the UI or conversation layer exists and who orchestrates the analytics workflow.
22
+
====
23
+
24
+
=== ThoughtSpot MCP Server
25
+
26
+
ThoughtSpot xref:mcp-integration.adoc[MCP Server] exposes governed analytics as MCP tools and resources to AI agents and clients. The MCP Server can be integrated with your MCP client, agent, LLM, or application UI, allowing your users to explore ThoughtSpot's agentic capabilities within the context of your application.
27
+
28
+
ThoughtSpot recommends using the MCP Server in these scenarios:
29
+
30
+
* When you want to plug ThoughtSpot into AI agents and clients that already support MCP, such as Claude, ChatGPT, Gemini, IDEs, and custom MCP clients.
31
+
* If you are building your own MCP-based chatbot or application, and want to call ThoughtSpot MCP tools behind a custom web experience.
32
+
33
+
For more information, see the xref:mcp-integration.adoc[MCP Server] documentation.
34
+
35
+
=== Embedding Spotter in your app
36
+
37
+
Developers can embed ThoughtSpot's xref:embed-spotter.adoc[conversational UI experience and its agentic capabilities] directly into their applications using the Visual Embed SDK.
38
+
39
+
If your business requires you to quickly integrate the Spotter experience into your applications and workflows, and provide a native AI search and analysis experience within your app, use this option.
40
+
41
+
For more information, see the xref:embed-ai-analytics.adoc[Embed AI Search and Analytics] and xref:embed-spotter.adoc[Spotter embed] documentation.
42
+
43
+
=== Spotter REST APIs
44
+
45
+
Spotter REST APIs provide programmatic access to Spotter conversation sessions, analytics, and agentic workflows. REST API clients use these APIs to send questions and receive structured answers, charts, or get relevant questions for a specific data model.
46
+
47
+
ThoughtSpot recommends using Spotter REST APIs in the following scenarios:
48
+
49
+
* When you want your application, agent, or orchestration logic to interact with ThoughtSpot programmatically, without requiring MCP.
50
+
* If you need to integrate ThoughtSpot responses and workflows with other systems.
51
+
* For fine-grained control over the user experience.
52
+
53
+
//* For fine-grained control over data context, query options, and prompt structure, and .
54
+
55
+
For more information, see the xref:spotter-apis.adoc[Spotter AI API] documentation.
56
+
57
+
== Additional resources
58
+
59
+
* For information about MCP, see the link:https://modelcontextprotocol.io[Model Context Protocol specification, window=_blank].
60
+
* For information about SDK libraries for embedding, see the link:https://github.com/thoughtspot/visual-embed-sdk[Visual Embed SDK GitHub repository, window=_blank].
61
+
* To view Spotter APIs, visit the +++<a href="{{navprefix}}/restV2-playground?apiResourceId=http%2Fgetting-started%2Fintroduction">REST API v2.0 Playground</a>+++ and navigate to the **AI** section.
You can now embed the Spotter 3 experience in your application and use features such as Auto mode for automatic data model selection, Chat history, and a new chat prompt interface.
71
+
You can now embed the Spotter 3 experience in your application and use features such as Auto mode for automatic data model selection, chat history, and a new chat prompt interface.
72
72
73
73
* To enable the new chat prompt interface, set `updatedSpotterChatPrompt` to `true`.
74
74
* To use Auto mode, set the `worksheetId` to `auto_mode`.
@@ -91,7 +91,7 @@ On Spotter embed deployments running version 26.2.0.cl or later, the *Add to Coa
91
91
92
92
Styling and grouping::
93
93
94
-
* The `isLiveboardStylingAndGrouping` attribute, used for enabling the Liveboard styling and grouping feature, is now replaced with `isLiveboardMasterpiecesEnabled`. While your existing configuration with the deprecated attribute, `isLiveboardStylingAndGrouping` continues to work, we recommend switching to the new configuration setting.
94
+
* The `isLiveboardStylingAndGrouping` attribute, used to enable the Liveboard styling and grouping feature, is now replaced with `isLiveboardMasterpiecesEnabled`. While your existing configuration with the deprecated `isLiveboardStylingAndGrouping` attribute continues to work, we recommend switching to the new configuration setting.
95
95
* The following action IDs are now available to show, disable, or hide the grouping menu actions on a Liveboard:
96
96
** `Action.MoveToGroup` for the **Move to Group** menu action.
97
97
** `Action.MoveOutOfGroup` for the **Move out of Group** menu action.
@@ -105,21 +105,21 @@ For more information, see link:https://docs.thoughtspot.com/cloud/latest/securit
105
105
+
106
106
The `showMaskedFilterChip` setting is also available in full application embedding.
To handle errors in the embedding workflows, the SDK includes the following features:
119
119
120
-
* `ErrorDetailsTypes` enum for categorizing error types, such as `API`, `VALIDATION_ERROR`, `NETWORK` error.
121
-
* `EmbedErrorCodes` enum with specific error codes for programmatic error handling
122
-
* `EmbedErrorDetailsEvent` interface for structured error event handling
120
+
* `ErrorDetailsTypes` enum for categorizing error types, such as `API`, `VALIDATION_ERROR`, and `NETWORK`.
121
+
* `EmbedErrorCodes` enum with specific error codes for programmatic error handling.
122
+
* `EmbedErrorDetailsEvent` interface for structured error event handling.
123
123
124
124
For more information, see link:https://developers.thoughtspot.com/docs/Enumeration_EmbedErrorCodes[EmbedErrorCodes] and link:https://developers.thoughtspot.com/docs/Enumeration_ErrorDetailsTypes[ErrorDetailsTypes].
125
125
|====
@@ -129,7 +129,7 @@ For more information, see link:https://developers.thoughtspot.com/docs/Enumerati
129
129
[width="100%" cols="1,4"]
130
130
|====
131
131
132
-
|[tag redBackground]#DEPRECATED# | **Use `minimumHeight` instead of `defaultHeight` ** +
132
+
|[tag redBackground]#DEPRECATED# | **Use `minimumHeight` instead of `defaultHeight`** +
133
133
134
134
The `defaultHeight` parameter is deprecated in Visual Embed SDK v1.44.2 and later.
135
135
To set the minimum height of the embed container for ThoughtSpot components such as a Liveboard, use the `minimumHeight` attribute instead.
@@ -144,7 +144,7 @@ Allows configuring which API calls to intercept.
144
144
* `interceptTimeout` +
145
145
Sets the timeout duration for handling interception.
146
146
* `isOnBeforeGetVizDataInterceptEnabled` +
147
-
When set to true, it enables the use of `EmbedEvent.OnBeforeGetVizDataIntercept` event to emit and intercept search execution calls initiated by the users and implement custom logic or workflow to allow or restrict search execution.
147
+
When set to true, it enables use of `EmbedEvent.OnBeforeGetVizDataIntercept` to emit and intercept search execution calls initiated by users and implement custom logic or workflows to allow or restrict search execution.
148
148
* `EmbedEvent.ApiIntercept` +
149
149
Emits when an API call matching the conditions defined in `interceptUrls` is detected.
150
150
@@ -157,7 +157,7 @@ For more information, see xref:api-intercept.adoc[Intercept API calls and search
0 commit comments