Introducing Bold BI MCP Tool: Setup, Features, and Use Cases

TL;DR:

• Bold BI's MCP Tool for AI Applications lets AI assistants like Claude, VS Code, and Cursor connect to Bold BI® Cloud or a publicly accessible hosted Bold BI site, so you can list dashboards, check permissions, inspect data sources, and export dashboards through plain-language requests instead of hand-written API calls.
• It exposes 51 MCP tools over a Streamable HTTP endpoint, authenticates with your existing Bold BI token, and is read-and-operate only: discovery, inspection, and export, not destructive changes.
• It works with Bold BI Cloud and with self-hosted Bold BI sites, as long as the target site is reachable from the MCP service.
• Every tool is granular and independently auditable, a deliberate tradeoff that trades one-shot convenience for governance visibility into exactly what an agent touched. 

Introduction

Imagine asking your AI assistant, “which dashboards does the Operations group have access to?” and getting a real, current answer instead of a guess. Today, that usually means opening an admin console, navigating to the right screen, and checking permissions manually, one dashboard and one group at a time.

Bold BI's MCP Tool makes this possible without leaving your AI assistant. It lets you securely connect compatible AI applications, such as Claude, VS Code, and Cursor, to Bold BI Cloud or a self-hosted Bold BI site, and interact with dashboard data through natural-language conversations.

One detail matter before you go further: the target Bold BI site, Cloud or self-hosted, must be reachable from the MCP service. If you're running Bold BI and want your AI assistant to actually understand your BI environment, this guide walks through setup, usage, evaluation criteria, and troubleshooting. Keep reading for detailed insights.

What Is MCP?

MCP is an open standard, created by Anthropic and released in November 2024, that lets AI applications connect to external tools and data sources through one common protocol instead of a custom integration per assistant. Without MCP, an AI assistant can only work with what it was trained on or what you paste into the chat, and it can't reach into your Bold BI site to check anything live.

That matters for BI specifically: it removes the "which AI assistant do we build for?" problem. Bold BI ships a set of MCP tools, so Claude, ChatGPT, Copilot, and custom agents can all connect to and work with Bold BI in the same way.

What Is Bold BI's MCP Tool?

The Bold BI MCP Tool is a Model Context Protocol (MCP) server that lets AI assistants and MCP-compatible applications interact with Bold BI through structured tools. Instead of calling Bold BI's REST API directly, a user asks a question in natural language, and the AI assistant invokes the matching MCP tool to retrieve information or perform an action on their behalf.

The Bold BI MCP Tool exposes 51 tools across dashboards, data sources, users, groups, schedules, slideshows, pinboards, items, categories, published content, and webhooks. Every tool call is scoped by the Bold BI token you provide, so an agent never sees more than that token already permits, and it only works if the target Bold BI site is reachable from the MCP service, as noted above.
To see how a single request actually travels through that system, it helps to name every piece involved.

Key Components of the Bold BI MCP Architecture

Five parts have to line up correctly for any of this to work.

Component	Description
MCP Client	An MCP-compatible application, such as Claude, VS Code, Cursor, or a custom AI agent, that supports Streamable HTTP transport.
MCP Service Endpoint	The hosted Bold BI MCP service that exposes the available tools (https://mcp.boldbi.com/api/mcp).
Authorization Header	Authenticates the request against the target Bold BI site.
ServiceURL Header	Identifies the target Bold BI site that should process the request.
Bold BI Site	Executes the requested operation and returns the result.

What You Can Do with Bold BI's MCP Tool

Once connected, your AI assistant can:

• Discover available dashboards and content.
• Retrieve dashboard, data source, user, and group information.
• Export dashboards and data sources.
• Review permissions and access history.
• Monitor schedules and governance information.
• Integrate Bold BI operations into AI-powered workflows.

All 51 tools fall into this pattern: discovery, inspection, and export.

Bold BI's MCP Tool vs. Traditional Workflow: Key Differences Explained

To understand the operational impact of MCP, it's useful to compare it with the traditional administrative workflow. The difference becomes apparent when reviewing permissions, access history, and dependencies across multiple assets.

Aspects	Bold BI MCP Tool	Traditional Workflow
Where you work	Inside your AI assistant	Requires switching to the Bold BI UI
Speed	One natural-language request	Navigate menus, click through screens
Auditing	Query access history directly	Manually cross-reference logs
Governance	Token-scoped, consistent, mapped to named tools	Depends on who's clicking around
Cross-referencing	Ask one question spanning multiple tools	Manually stitch data source, dashboard, and permission views together
Client flexibility	The same MCP Tool works across Claude, VS Code, Cursor, and other MCP-compatible clients	Each admin console is a separate UI, with no equivalent for AI clients

How the MCP Tool Works

The Bold BI MCP Tool acts as a bridge between an MCP-compatible client and a Bold BI site. The sequence runs like this:

1. The user submits a request through an AI assistant or MCP-compatible client.
2. The MCP client sends the request to the Bold BI MCP service.
3. The MCP service authenticates the request using the token supplied in the Authorization header.
4. The MCP service routes the request to the target Bold BI site specified in the ServiceURL header.
5. The Bold BI site API executes the requested operation.
6. The result is returned to the MCP service.
7. The MCP service returns the response to the MCP client.
8. The AI assistant presents the result to the user.

Access is never broader than what the underlying token already allows. And the request only completes at all if your Bold BI site, Cloud or self-hosted, is reachable from the MCP service in the first place.

How to Set Up the MCP Connection in Bold BI

Before you start your connection, you need:

Bold BI requirements

• A reachable Bold BI site Caller URL.
• A valid Bold BI user account or service account.
• A valid Bold BI API key, generated from your Bold BI user profile.
• Permissions required for the operations you want to perform.

Required connection information

• Bold BI MCP endpoint URL.
• Bold BI site API URL.
• A valid Bold BI API key.

Network requirements

• Outbound HTTPS access to https://mcp.boldbi.com/api/mcp.
• Firewall rules that don't block access to the MCP endpoint.
• Proxy configurations that allow traffic to the hosted service.
• The target Bold BI site must be reachable from the MCP service.

MCP client requirements

Your client must support:

• Model Context Protocol (MCP).
• Streamable HTTP transport.
• Custom HTTP headers.

Once these prerequisites are in place, you can connect an MCP-compatible client to the Bold BI MCP Tool by following the steps below.

Step 1: Get the MCP Service Endpoint

Every client points at the same hosted endpoint:

https://mcp.boldbi.com/api/mcp

Step 2: Generate Your API Key

The MCP service forwards the Authorization header it receives directly to the Bold BI REST API. Authentication uses a single Bold BI API key, there's no separate Bearer or Basic token to choose between.

Form	Example	Notes
API key	Authorization: <bold-bi-api-key>	Generate this from your Bold BI user profile.

Generate the key from your Bold BI profile page, then pass it as-is in the Authorization header, no prefix required. For step-by-step instructions, see Generate a Bold BI API Key.

Step 3: Set the Required Headers

Header	Required	Description
Authorization	Yes	Authentication token used to access Bold BI.
ServiceURL	Yes	API URL of the target Bold BI site.

 Step 4: Configure the Client

Every tool call that reaches Bold BI must include the Authorization and ServiceURL headers described above. Remember that ServiceURL needs to point to Bold BI Cloud or to a self-hosted site that's reachable from the MCP service.

Example configuration using your API key 

{

"servers": {

"boldbi": {

"type": "http",

"url": "https://mcp.boldbi.com/api/mcp",

"headers": {

"Authorization": "${BOLDBI_TOKEN}",

"ServiceURL": "http://your-boldbi-host/bi/api/site/site1"

}

}

}

}

Set BOLDBI_TOKEN to the API key from your Bold BI profile. Replace http://your-boldbi-host/bi/api/site/site1 with your own Bold BI site's API base URL.

Step 5: Verify the Connection

Ask your client to run tools/list. If the headers are right, you'll see all 51 Bold BI tools come back, and you're ready to start calling them with plain-language prompts.

Use Cases and Benefits of Bold BI MCP Tool for AI-Assisted BI Workflows

One of the biggest advantages of Bold BI's MCP Tool Support is its ability to combine information from multiple areas of your BI environment and return a single, contextual response through natural-language requests. The following examples show how different teams can use MCP to simplify common business and administrative tasks.

Use Case: Auditing Sales Analysis Dashboard Dependencies Before a Data Source Change

Scenario

A BI admin is planning to modify the "Sales Analysis Data" connection, maybe a credential rotation, a schema change, or a full migration to a new database. Before touching anything, they need to answer a deceptively hard question: what actually depends on this data source, and who can see it?

The problem

In a traditional workflow, this means opening the Bold BI admin console, scanning the full list of dashboards, opening each one to check its data source, then separately checking the users and groups with access, and finally exporting a copy for the change-management ticket. For a site with dozens of dashboards, that's a lot of manual clicking, and it's easy to miss a dependency.

How Bold BI's MCP Tool solves it

Instead, the admin just talks to their AI assistant:

• "List the all dashboard details" → the assistant calls dashboard_list and returns every dashboard on the site in seconds.
• "List the all datasource details" → datasource_list surfaces every connected data source, including the one about to change.
• "Share the details of sales analysis datasource connected dashboard details" → datasource_linked_dashboards_list immediately confirms that only one dashboard, "Sales Analysis Dashboard," depends on it, with its ID, category, and last-modified date. No dependency is missed because the tool queries the relationship directly instead of relying on memory or spreadsheets.
• "Show the widgets details available on [the] dashboard" → dashboard_widgets_list breaks the dashboard down widget by widget (type, subtype, grid position, widget ID), so the admin knows exactly what will be affected if the underlying data shape changes.
• "Export the sales analysis datasource in PDF format" → dashboard_export generates a PDF snapshot for the change ticket, with an explicit approval prompt before anything is exported, keeping a human in the loop.
• "List the user details" and "List the groups details" → user_list and group_list pull who currently has access, so the admin can flag stakeholders before the change goes live.

The result

What used to be a 20-minute cross-referencing exercise across multiple admin screens becomes a single conversational thread, fully answered in under two minutes, with every step (list, lookup, export) individually logged and auditable. The admin walks into their change-management review with a complete dependency map, an exported artifact, and an access list, without ever leaving their AI assistant.

How Bold BI's 51 MCP Tools are Organized

The MCP tools catalog is organized into 11 functional groups that mirror major areas of the Bold BI platform, covering discovery, inspection, and export across dashboards, data, users, and governance.

Category	Tools	What it covers	Representative tools
Dashboard	15	Full dashboard lifecycle — listing, permissions, exports, access history	dashboard_list, dashboard_permissions_get, dashboard_access_history_get
Dashboard settings	2	Global dashboard configuration and deployment info	dashboard_settings_get, deployment_dashboard_list
Category	2	How dashboards are organized into categories	category_list, category_get
Data source	6	Structure, dependencies, exports, and activity for data sources	datasource_get, datasource_linked_dashboards_list, datasource_activity_get
User	8	Individual users, group memberships, permissions, bulk import status	user_get, user_permissions_get, user_import_job_status_get
Group	5	Groups, members, attributes, and permissions	group_list, group_users_list, group_permissions_get
Item and pinboard	6	Items and pinboards, including naming-conflict checks	item_list, pinboard_list, item_name_exists_check
Published content	2	What's published externally and where	published_item_list, external_publish_site_list
Schedule	2	Scheduled runs and their configuration	schedule_list, schedule_get
Slideshow	2	Slideshows configured on the site	slideshow_list, slideshow_get
Webhook	1	Webhooks configured on the site	webhook_list

 

For the complete tool-by-tool reference, including every parameter and return type, see the Bold BI MCP Tool documentation.

Evaluating MCP Implementations: Approaches and What to Check

Not every MCP implementation is built the same way, and the differences affect security, governance, maintainability, and long-term reliability just as much as convenience. MCP solutions generally fall into one of three approaches, depending on the level of transparency, control, and customization an organization requires.

Approach	How it works	Best fit	Where Bold BI's MCP tool fit
Granular, read-and-operate tools	Many narrow, named tools; the AI client chains calls itself	Teams that need auditability, explicit permissions mapping, and a verifiable no-write guarantee	This is Bold BI's approach, exposing 51 tools across dashboard discovery, inspection, governance, auditing, and export operations.
Single-instruction orchestration	One broad tool accepts a natural-language instruction; an internal agent layer resolves it into multiple actions	Teams that prioritize convenience and rapid task completion through a single request	Not Bold BI's model. By design, every action stays independently visible
Build-your-own	No hosted tools; you build and host an MCP integration yourself against the vendor's REST API	Teams with engineering capacity who need deep customization beyond what any hosted tool set exposes	Typically unnecessary for Bold BI use cases because the hosted MCP tool already covers dashboards, data sources, users, groups, governance, schedules, and content discovery

Beyond choosing an approach, confirm who's actually responsible for keeping the integration alive. A community-built connector can break silently when the underlying API changes, with no guaranteed fix timeline. Bold BI's MCP Tool Support is officially built and maintained by Bold BI, documented at a single hosted endpoint, so updates and fixes come from the vendor itself rather than an unmaintained third-party project.

Common Issues and How to Fix Them

Most connection and configuration issues can be traced back to a few common causes. The table below highlights the most frequently encountered problems and the steps to resolve them.

Problem	Likely cause	How to fix it
Hosted MCP service cannot be reached	Incorrect MCP service URL or network block	Use https://mcp.boldbi.com/api/mcp and verify outbound HTTPS access.
Tool call says a token is required	Missing Authorization header	Send the API key in the Authorization header: Authorization: <bold-bi-api-key>.
Tool call returns unauthorized	Expired or insufficient Bold BI token	Refresh the token or use an account with the required permission.
Tool call fails to reach the correct site	Missing or incorrect ServiceURL header	Include the ServiceURL header with the correct Bold BI site API URL.

Best Practices for Using Bold BI MCP Tool

To get the most value from Bold BI's MCP Tool Support, follow these best practices:

• Secure authentication tokens: Store tokens in environment variables or a secure secret-management solution instead of hard-coding them into applications or configuration files.
• Protect sensitive credentials: Never commit authentication tokens, passwords, or other credentials to source control repositories.
• Use dedicated service accounts: Create separate service accounts for automation and AI-driven workflows to simplify management and auditing.
• Apply least-privilege access: Grant only the permissions required for the operations the AI assistant needs to perform.
• Use secure connections: Ensure all communication with the MCP service and your Bold BI environment is conducted over HTTPS.
• Avoid sensitive information in prompts: Do not include confidential business data, passwords, or credentials when interacting with AI assistants.
• Rotate tokens regularly: Periodically refresh authentication tokens according to your organization's security policies.
• Be specific in your requests: Clearly identify the dashboard, data source, user, group, or item you want the AI assistant to work with.
• State your objective clearly: Explain whether you're reviewing permissions, checking access history, analyzing dependencies, or exporting content.
• Combine related questions: Group related requests together so the AI assistant can automatically chain the appropriate MCP tools and provide a more complete response.
• Define expected outputs: Specify details such as export formats, date ranges, or the exact information you want returned to receive more relevant results.

[elementor-template id="150021"]

Want to Explore Your Bold BI Environment Through AI?

Bold BI® MCP Tool extends your BI environment to MCP-compatible AI applications through a secure, governed, and standardized interface. This approach simplifies information discovery and analysis while preserving the permissions and controls already defined within Bold BI.

Ready to see how AI can interact with your analytics environment? Connect an MCP-compatible client to Bold BI and start exploring your dashboards, data sources, and governance insights through natural-language requests.

If you're already using Bold BI, this capability is available today. All you need to do is connect your client and run the tool/list to explore the available MCP tool. Ready to get started? Sign up for a free trial or request a personalized demo.'

Related Resources:

• Bold BI MCP Server Setup Guide for AI Assistants | Bold BI
• Bold BI Documentation Is Now Available as an MCP Server

Frequently asked questions

1. 1. 1.

      What is the Model Context Protocol (MCP)?

      An open standard, created by Anthropic and released in November 2024, that lets AI applications connect to external tools and data sources through one common protocol instead of one-off integrations.
   2. 2.

      What URL do Bold BI's MCP tool use?

      https://mcp.boldbi.com/api/mcp. Configure your client with this URL, an Authorization header, and a ServiceURL header pointing to your Bold BI site.
   3. 3.

      Does this work with any Bold BI site?

      It works with Bold BI Cloud and with self-hosted Bold BI sites, provided the target site is reachable from the MCP service. 
   4. 4.

      Which clients work with it?

      Any MCP-compatible client that supports Streamable HTTP: Claude, VS Code, Cursor, and custom agents built against the MCP spec.
   5. 5.

      How many tools does it expose, and why so many?

      51, across dashboards, dashboard settings, categories, data sources, users, groups, items and pinboards, published content, schedules, slideshows, and webhooks. The count is deliberately granular rather than consolidated into a few broad instructions, so every action an agent takes maps to a specific, auditable Bold BI operation.

   6. 6.

      Can it modify or delete dashboards?

      No. It's read-and-operate: discovery, inspection, and export only.
   7. 7.

      Do I need a special token?

      Just a valid Bold BI API key, generated from your Bold BI user profile. The MCP service forwards it to the Bold BI REST API as-is.