The Cube MCP Server lets you connect your governed financial data directly to AI platforms like Claude and ChatGPT. Instead of exporting data and pasting it into AI chats, the MCP Server provides a live, secure connection that respects your Cube permissions on every request.
Overview
With the Cube MCP Server, you can:
- Query live Cube data from Claude, ChatGPT, and other MCP-compatible AI tools: actuals, budgets, forecasts, variances, and more
- Ask an AI to generate board decks, run variance analysis, or build financial summaries using your real Cube data
- Read your planning Tables, such as headcount rosters and asset registers, at line-item detail
- See your source connections and import history, and refresh actuals on demand rather than waiting for the next scheduled import
- Review your close and planning workflows and the tasks inside them
- Have an AI write updated figures back into a scenario, populate a planning Table, create a scenario, build and update Boards, draft workflows and tasks, and create tags, when you grant write access
- Drill into transaction-level detail and dimension breakdowns directly from an AI chat
- Maintain full security: Cube's role-based access control (RBAC) security model enforces your user permissions on every AI request. If you can't see it in Cube, the AI can't see it either
For every call an AI tool can make, what write access allows, and what each call costs in AI credits, see What AI Tools Can Do Through the Cube MCP Server.
Note: For endpoint details and technical reference, see the Cube MCP Server documentation.
How it works
The Model Context Protocol (MCP) is an open standard that lets AI tools connect to external data sources. Cube's MCP Server acts as a secure gateway between your AI tool and your Cube data.
When you (or an AI agent acting on your behalf) send a query through MCP, Cube validates your identity, checks your Data Access Scope, and returns only the data you are authorized to see. The AI tool receives governed, calculated, multi-currency-normalized financials, not raw exports. The same governed connection also lets the AI build live Boards, populate planning Tables, and draft workflows in your workspace from that data.
This means:
- Every response uses your Cube Source of Truth: consolidated, calculated, and current
- Your Cube business rules, formulas, and multi-currency logic are applied before data reaches the AI
- No data leaves Cube without passing through your organization's permission model
What the AI can do
Once connected, an AI tool can read your Cube data within your Data Access Scope and, if you have write access, build Boards and Tables, write data back to scenarios opened for AI, and draft workflows and tags. Access comes in two permission levels:
- Use MCP (Read Only): the AI can query and read your data. It cannot change anything.
- Use MCP (Write): the AI can also write data back and build in Cube, subject to write-back rules that keep it from overwriting scenarios you rely on.
Data calls and writes use your company's AI credits. For the full list of calls, the write-back rules, and credit costs, see What AI Tools Can Do Through the Cube MCP Server.
Prerequisites
Before connecting, confirm the following:
- You have a Cube account with an active login.
- You have access to an AI tool that supports MCP connectors. Note that MCP functionality may require a specific subscription tier depending on the tool:
- Claude: Cube is listed in the Claude Connectors Directory. Connector availability depends on your Claude plan. Check Anthropic's plan details for the latest information.
- ChatGPT: MCP connectors require a Plus, Pro, Team, Business, Enterprise, or Edu plan. Developer Mode must be enabled by a workspace admin for Business and Enterprise plans. Check OpenAI's plan details for the latest information.
- Other tools: Refer to your tool's documentation for MCP support and plan requirements.
- To write data back, build or update Boards, or create tags, your Cube user needs Use MCP (Write). To query and read, Use MCP (Read Only) is enough.
- Your company has AI credits available. Data calls and writes use credits from your company's shared pool. See How AI Credits Work.
Connect to Claude
Enterprise and Team plans: admin setup first
If your organization is on a Claude Enterprise or Team plan, an Owner or Primary Owner needs to add Cube for the organization before members can connect. This is a one-time setup.
An Owner or Primary Owner should:
- Sign in to Claude and go to Organization settings > Connectors.
- Select + Add and go to "All Available".
-
Search for Cube Software, then select Add.
Cube now appears in your organization's connector list.
Owners can also control which Cube tool calls are available to members. In Organization settings > Connectors, you can allow Cube to read data while restricting the tools that build or update Boards, applied across everyone in the organization. This is optional. If you want members to use the full set of Cube capabilities, no change is needed.
Connect your account
Once Cube is available (immediately on Pro plans, or after an Owner adds it on Team and Enterprise plans), each user connects their own account:
- Open the Cube listing at https://claude.ai/directory/cube-software, or go to https://claude.ai/customize/connectors and search for Cube Software.
-
Select Connect.
- Sign in with your Cube credentials when prompted.
Once connected, Cube appears in your list of connectors.
Note: Prefer to add Cube manually? You can still add it as a custom connector. Use Add custom connector (Owners: Organization settings > Connectors > Add), enter Cube as the name and
https://mcp.cubesoftware.com/as the URL, then sign in with your Cube credentials.
Start querying:
Open a new conversation in Claude and ask a question about your financial data. Examples:
- "Build me a board deck showing this quarter's performance. Include revenue by quarter, gross margin trends, opex breakdown by department, and headcount growth. Compare actuals to budget."
- "Create a summary comparing Actuals vs. Forecast for Revenue by entity for the last 12 months."
- "If we cut Sales & Marketing spend by 15% in the back half of the year, how would that change our Net Income? Compare to what we actually spent."
Claude will call the Cube MCP Server to retrieve your data, apply your permissions, and return results grounded in your Source of Truth.
Connect to ChatGPT
ChatGPT supports custom MCP connectors through its Apps & Connectors settings. This feature requires a ChatGPT Plus, Pro, Team, Business, Enterprise, or Edu plan.
You can add Cube directly as a ChatGPT app instead of creating a custom connector. This is the fastest way to get started.
Requirements: ChatGPT Plus, Pro, Team, Business, Enterprise, or Edu plan.
- Open this link: Add Cube to ChatGPT
- ChatGPT will open the Cube app page. Click Connect.
- Review the permissions summary, then click Continue to Cube.
- Sign in with your Cube credentials to authorize the connection.
- Once connected, Cube will appear in your list of apps under Settings > Apps & Connectors.
- To start querying, open a new chat, click the + button (or Connected apps) in the chat tools menu, enable Cube, and ask a question about your financial data.
If you prefer to configure the connection manually, or if the app link above doesn't work for your ChatGPT plan, follow the steps below.
Step 1: Enable Developer Mode
- In ChatGPT, go to Settings > Apps & Connectors > Advanced settings.
- Enable the Developer Mode toggle.
Step 2: Add the Cube connector
- Return to Settings > Apps & Connectors.
- Click Create (or + Create New Connector).
- Enter the following:
- Name: Cube
- Description: Query governed financial data from Cube, including actuals, budgets, forecasts, variances, and transaction-level detail.
-
MCP Server URL:
https://mcp.cubesoftware.com/ - Authentication: OAuth
- Check the box to confirm you trust this application.
- Click Create.
- You will be redirected to Cube's login page. Sign in with your Cube credentials to authorize the connection.
- Once connected, Cube will appear in your list of connectors under Apps & Connectors.
Step 3: Start querying
- Open a new chat in ChatGPT.
- Click the + button (or Connected apps) in the chat tools menu and enable the Cube connector.
- Ask a question about your financial data. Examples:
- "Use deep research to analyze our company's financial performance in the context of current market conditions. Pull our revenue, margins, and opex from Cube, then research comparable SaaS benchmarks and write a board-ready market positioning report."
- "Pull our quarterly P&L from Cube for 2024 and 2025 and open it in Canvas. Build a variance analysis with charts showing actuals vs budget by department, and let me edit the commentary inline before exporting to PDF."
- "Connect to Cube and pull our current headcount, personnel costs, and revenue by department. Then model three scenarios: (1) 10% RIF in Sales, (2) 20% Engineering hiring ramp, (3) flat headcount with 5% raises. Show the net income impact of each."
ChatGPT will display a confirmation prompt before executing any tool call, so you can review what data is being requested.
Troubleshooting an "Unknown Tool" related error:
- Click your name in the lower left-hand corner of the screen.
- Select Settings from the menu.
- Select Apps from the new window.
- Select Cube from your list of apps.
- You see something that says Information. To the right of that, there should be a Refresh button. Click that button.
Connect Cube to Microsoft Copilot
Extend a Microsoft Copilot Studio agent with live access to your Cube data, using the same governed permissions your users already have in Cube.
Cube connects to Copilot Studio as a custom MCP server and uses OAuth 2.0 with dynamic discovery. Copilot Studio finds Cube's OAuth settings automatically, so you do not need a Client ID, Client Secret, or scopes from Cube. Copilot Studio supports the Streamable HTTP transport, which Cube's MCP server uses.
Requirements
- A Microsoft 365 Copilot license or a Copilot Studio User License, assigned to the user who builds the agent. Your Microsoft 365 admin may need to assign it.
- Permission to create and publish agents in Copilot Studio.
- A Cube user with Use MCP (Read Only) at minimum, or Use MCP (Write) for write-back.
Add Cube to your Copilot agent
- Open Copilot Studio and select the agent you want to extend. Note: Use a standard agent. Do not use the built-in Copilot for PowerApps agent. It has no owner, so you cannot publish or share it with other users, and an owner cannot be added later. If you started there, create a new standard agent and add Cube to it.
- Select Settings, then go to the Generative AI section.
- Under Orchestration, select Yes for Use generative AI orchestration for your agent's responses, then select Save.
- Go to the Tools tab and select Add a tool.
- Under Create new, select Model Context Protocol.
- Enter the following:
- Server name: Cube
- Server description: Query governed Cube data, including actuals, budgets, forecasts, scenarios, workflows, and Tables, with your Cube permissions.
-
Server URL: https://mcp.cubesoftware.com/
Note: Enter the URL exactly as shown, with no /mcp at the end. After you save, Copilot Studio does not display the values you entered, so confirm them before you continue.
- Under Authentication, select OAuth 2.0.
- Select Dynamic discovery.
- Select Create.
Connect and test
- On the Tools tab, open the Cube tool.
- Under Connection, open the dropdown and select Create new connection.
- Sign in with your Cube credentials to authorize the integration.
- Select Add and configure.
- Confirm that the Cube tools appear in the agent's tool list.
- In the test pane, ask a question such as "What are our total revenue actuals for the last full quarter?"
The integration is working when the Cube tools are listed and the agent answers with your Cube data.
Troubleshooting
The agent works for you but not for other users. Confirm the agent is a standard agent with an owner, not the built-in Copilot for PowerApps agent. If it is, create a new standard agent and add Cube again.
Your IT team needs the OAuth endpoints. Cube publishes its OAuth 2.0 metadata at https://api.cubesoftware.com/.well-known/oauth-authorization-server. Copilot Studio reads it automatically when you select Dynamic discovery.
For any additional questions, contact support@cubesoftware.com for additional assistance.
Governance notes for admins
- Every Cube MCP request is validated against the signed-in user's Cube permissions. Copilot cannot see or write data that the user could not see or write in Cube directly.
- To restrict Copilot to read-only access, assign users the Use MCP (Read Only) permission in Cube.
Connect Cube to Gemini Enterprise
Gemini Enterprise supports custom MCP servers through its Data Stores feature, letting your users query Cube from the Gemini Enterprise app with the same governed permissions they have in Cube.
Cube connects to Gemini Enterprise as a Custom MCP Server data store, using OAuth 2.0 for per-user authentication. Gemini Enterprise only supports the StreamableHTTP transport, which Cube's MCP endpoint uses.
Requirements
- A Gemini Enterprise subscription with access to the Google Cloud console.
- A Google Cloud project where you can create Data Stores.
- The Discovery Engine Admin role (or at minimum Discovery Engine Editor) granted to the user setting up the data store.
- If your project uses VPC Service Controls or has project enforcement enabled, an administrator must add custom_mcp to the allowed data sources and allow the mcp.cubesoftware.com domain.
- A Cube user account with Use MCP (Read Only) at minimum, or Use MCP (Write) if you want your agent to write back to scenarios, build Boards, or create tags.
Admin-only setup. Custom MCP data stores in Gemini Enterprise are configured at the project level by a Google Cloud administrator. Individual end users cannot add Cube on their own; once an admin sets up the data store and enables actions, each user authorizes their own Cube session inside the Gemini Enterprise chat.
Set up the Cube data store in Gemini Enterprise
- In the Google Cloud console, open Gemini Enterprise.
- In the navigation menu, click Data stores, then click Create data store.
- In the Select a data source page, search for Custom MCP Server and click Add MCP server.
- In the MCP Server Configuration page, enter:
- Authentication settings: select OAuth 2.0
- MCP Server URL: https://mcp.cubesoftware.com/mcp
- Enter the OAuth values provided by Cube. Customers who need OAuth credentials should contact support@cubesoftware.com.
- Check “Enable PKCE Support”
- Check “Use HTTP Basic Authentication”
- Click Verify Auth, then Continue.
- In Configure your data connector, select a Location (Multi-region) and enter a name for the data store, for example Cube.
- Click Create. Gemini Enterprise creates the data store and returns you to the Data stores page.
Enable Cube actions
By default, no actions from a custom MCP server are enabled. You have to turn them on before users can call Cube tools.
- From the Data stores list, open the Cube data store you just created.
- In the side menu, click Actions.
- Click Reload custom actions.
- Select the Cube actions you want to make available to your users, then click Enable actions.
Connect and authorize as an end user
Once an administrator has enabled Cube actions on the data store, each user needs to authorize their own Cube session the first time they use it.
- Open the Gemini Enterprise chat.
- At the bottom of the chat window, find your Connected Data Stores.
- Locate the Cube connector and toggle it ON.
- When prompted, sign in with your Cube credentials to grant Gemini Enterprise permission to act on your behalf.
- Once the authorization succeeds, the broken-chain icon next to the connector clears, and you can start asking questions about your financial data.
Because Gemini Enterprise validates every request against your Cube user permissions, agents only see the data your Cube role would let you see, and write actions require Use MCP (Write) on your Cube user.
Connect to other MCP-compatible tools
The Cube MCP Server works with any tool that supports the Model Context Protocol. The general pattern is:
- Open your tool's MCP or Connectors settings.
- Add a new connector with the URL
https://mcp.cubesoftware.com/. - Authenticate with your Cube credentials when prompted.
Refer to your specific tool's MCP documentation for exact configuration steps. For Cube-side details, such as the server endpoint and authentication, see the Cube MCP Server documentation.
Things to keep in mind
- Every MCP request is validated against your Cube user permissions. There are no master keys or backdoors.
- The MCP Server delivers governed, calculated data. Cube's rules engine, multi-currency logic, and multi-source consolidation are applied before data reaches any AI tool.
- If your Data Access is modified in Cube, the change takes effect on the next MCP request in real time.
- AI tools may interpret or summarize data in their responses. Always verify AI-generated content against your Cube Source of Truth before using it in formal reporting or decision-making.
- Data calls and writes use your company's AI credits. Check the MCP tier on the Usage page to see how much your AI tools are using.
- See how to get started with example prompts here.
Troubleshooting
- If your connection has stopped working, or is presenting unexpected errors, try to disconnect and reconnect to the server. If that doesn't work, contact our Support team.
- Getting a 401 Invalid Token error when you really don't think that you should? Check to be sure that your Scope is set to "read write introspection".
- If your original configuration was set to "Read Only" initially, you may need to recreate the configuration to allow for reading, writing, and introspection in addition to updating the Scope setting.
- If an AI reports that it cannot write to a scenario, the target scenario is almost always not open to AI write-back. See What AI Tools Can Do Through the Cube MCP Server for how to open a scenario for AI.
- If the AI reports that "Your company has run out of Cube AI credits," the connection is fine. Your company's balance is too low for that call. See What Happens When You Run Out of AI Credits.