ThoughtSpot’s SpotterCode is an AI-powered coding assistant that streamlines and speeds up the process of embedding ThoughtSpot content and integrating ThoughtSpot REST APIs. SpotterCode is available in your IDE, the Visual Embed Playground, and on the ThoughtSpot developer documentation site.
SpotterCode connects to a ThoughtSpot-hosted MCP server and provides AI-assisted coding and documentation capabilities across multiple surfaces. It gives developers direct access to ThoughtSpot SDKs, REST API documentation, code samples, and developer guides, whether they are working in an AI-native IDE, experimenting in the Visual Embed Playground, or browsing the ThoughtSpot developer documentation site.
SpotterCode is available in the following surfaces:
| Surface | Description |
|---|---|
IDE | Connects your AI-native IDE to the SpotterCode MCP server. The AI agent in your IDE can generate embed code, execute authenticated ThoughtSpot API workflows, and look up documentation without leaving your development environment. SpotterCode supports the following IDEs:
For more information, see Integrate SpotterCode with your IDE. |
Visual Embed Playground | A built-in AI coding assistant on the Playground page in your ThoughtSpot instance. SpotterCode appears as a slideout panel and generates embed code for the component you are configuring, using quick-starter prompts or custom instructions. See SpotterCode in the Visual Embed Playground. |
Developer documentation site | The SpotterCode chat assistant is available on the ThoughtSpot developer documentation site ( |
SpotterCode is designed for developers and technical teams who embed ThoughtSpot content or integrate ThoughtSpot REST APIs into their applications.
IDE users: Developers using AI-native IDEs such as Cursor, Visual Studio Code with GitHub Copilot, or Claude Code who want context-aware code generation and direct API execution from their development environment.
Playground users: Developers exploring embedding options in the Visual Embed Playground who want to generate and iterate on embed code quickly without writing boilerplate manually.
Documentation site visitors: Developers reading the ThoughtSpot developer documentation who want to ask questions about embedding, APIs, or SDK configuration in context.
SpotterCode is useful at every stage of an embedded ThoughtSpot project:
Setting up a new embedded project
Use the authenticated /mcp endpoint so the IDE agent can configure your ThoughtSpot instance directly. You can generate a trusted authentication secret key, add your application domain to CORS and CSP allowlists, and retrieve Liveboard and Answer object IDs, all without leaving your IDE.
Writing and reviewing embed code
SpotterCode gives your IDE agent direct access to Visual Embed SDK and REST API documentation, code samples, and developer guides. The agent can generate context-aware embed code tailored to your project structure and provide direct access to the SDK reference.
Integrating REST API workflows
Access up-to-date REST API specifications, request and response formats, authentication flows, and TypeScript and Java SDK references to accelerate API integration.
Looking up documentation
SpotterCode empowers your IDE with documentation lookup capabilities and provides direct access to authoritative information on embedding ThoughtSpot or integrating REST API workflows in your development projects.
|
Note
|
SpotterCode is an acceleration tool designed to help developers streamline the process of integrating ThoughtSpot into their projects. It does not replace the Visual Embed SDK or the foundational knowledge required for embedding or application integration. |
SpotterCode is an add-on tool available with the ThoughtSpot Analytics and ThoughtSpot Embedded offerings. If you have a ThoughtSpot Analytics license with an active ThoughtSpot Embedded subscription, you can integrate SpotterCode using the SpotterCode MCP server URL in your coding application.
ThoughtSpot provides the following MCP Server endpoints for SpotterCode:
| Endpoint | URL | Available tools |
|---|---|---|
Authenticated endpoint |
|
|
Documentation endpoint |
|
|
SpotterCode provides the following skills to the AI agent in your IDE. Authenticated skills require a valid ThoughtSpot session established through the SpotterCode authentication flow.
get-rest-api-reference
Provides REST API specifications, endpoints, request/response formats, authentication flows, CRUD operations, and SDKs for TypeScript and Java.
get-developer-docs-reference
Provides access to documentation on embedding, UI customization, deployment, security, and best practices.
execute-thoughtspot-code
Executes authenticated ThoughtSpot API workflows directly from your IDE agent. This skill allows your IDE agent to perform common operations that can be done via public REST APIs.
Use the following guidance to decide which endpoint to configure in your IDE:
Use /mcp (authenticated) if:
You want the AI agent to execute ThoughtSpot API workflows from your IDE, for example, to generate a trusted authentication secret key, configure your application domain in CORS/CSP allowlists, or retrieve Liveboard and Answer object IDs from your instance.
You want your IDE to use the full set of SpotterCode skills, including run-ts-workflow.
You are building or configuring a ThoughtSpot embedded application and want end-to-end setup assistance without leaving the IDE.
Use /mcp/docs (unauthenticated) if:
You only need the AI agent to look up documentation, REST API reference material, and code samples.
You are working in an environment where connecting to a ThoughtSpot instance is not appropriate. For example, a shared CI environment, a read-only developer workstation, or a demonstration setup.
You do not have a ThoughtSpot Embedded subscription or instance credentials available.
|
Note
|
The IDE configuration steps in this guide use the authenticated |
SpotterCode is designed to operate within ThoughtSpot’s existing security and governance model. The following table shows what data SpotterCode sends to its hosted MCP server and how your ThoughtSpot access controls apply across all SpotterCode surfaces.
| Surface | Authentication required | RLS/CLS enforced | Connects to ThoughtSpot instance | Data sent to MCP server |
|---|---|---|---|---|
IDE | Yes. OAuth/SAML or bearer token | Yes | Yes | User queries, the IDE agent’s tool call requests, and the authentication context required to execute code on the user’s behalf. SpotterCode does not send your source code, local file contents, or project structure to the MCP server. |
Visual Embed Playground | Yes. Requires an active ThoughtSpot session. | Yes | Yes | User queries sent within the Playground session. The Playground operates within your authenticated ThoughtSpot session. No additional credentials are transmitted separately. |
Developer documentation site | No | N/A | No | User queries and page context. No authentication or instance data is required. The documentation site surface is publicly accessible. |
SpotterCode does not bypass or modify ThoughtSpot’s existing access control model. When SpotterCode executes operations against your ThoughtSpot instance via the execute-thoughtspot-code tool, it acts as the authenticated user whose session token is in use, and can perform both read and write operations permitted to that user. The following ThoughtSpot security features remain in effect:
Because code executes as the authenticated user, write operations, such as create, update, delete, and share, generate the same audit log entries as equivalent REST API calls made directly by that user. For details on the events captured and how to retrieve them, see Audit logs and Audit logs API.
|
Note
|
SpotterCode performs operations as the authenticated user. Grant SpotterCode access only to accounts with the minimum permissions required for the intended use case. |
Responses from SpotterCode are determined by the features and parameters currently supported in the Visual Embed SDK, REST API, and official ThoughtSpot developer documentation. SpotterCode cannot generate code or solutions that rely on unsupported or undocumented features.
SpotterCode can generate code samples for the most common use cases. Scenarios that require advanced customization or highly specialized workflows may require manual intervention or additional coding beyond what SpotterCode provides.
SpotterCode does not influence how the Spotter feature in your ThoughtSpot embed infers semantic modeling. SpotterCode is not intended for querying data models or interpreting definitions such as measures and attributes in your metadata objects.
The behavior of the IDE agent, including tool selection and reasoning, is not controlled by SpotterCode.
