MCP integrations

MCP (Model Context Protocol) lets FastBI work with compatible external data sources and AI clients.

Add an MCP data source

Open Connections → Add connection → MCP integration.
In Access configuration, enter Server URL and select Continue. FastBI checks the server’s authentication requirements. Continue for a public server, enter a Bearer token, or select Authorize for OAuth. If requested, supply the public client ID issued by the provider for FastBI.
In Connector choices, review the discovered tools and resources. Select Use these tools and resources, then Continue. Discovery and authorization do not create a saved connection.
In Name your connection, enter Connection name and Slug, then select Save & test. The slug starts with a lowercase letter and uses lowercase letters, numbers, or hyphens.
After connection, browse the available sources in Enrichments and use the relevant source in your analysis.

The data available depends on what the MCP server exposes and the permissions granted to FastBI. If authorization expires or is cancelled, reopen the connection setup and retry.

Connect an AI client to FastBI

Follow Connect your AI chats for the complete OAuth flow, one-off questions, scheduled report retrieval, and access renewal.

Open User settings → Connect your AI with MCP and copy the MCP server URL.

For an AI client with OAuth support, add that URL in the client and choose its connection action. Sign in to FastBI, select your organization, and review and approve the requested access.

For a client that uses a bearer token, enter a connection name under Manual access tokens, choose the Allowed features, and select Create access token. Copy the token when shown and configure it in your client’s bearer-token field. Manual tokens expire after 30 days. Use Revoke in User settings to remove a token or a connected AI app’s access.

Requested data is sent to the AI client. Analysis and SQL queries use your usual FastBI credits, and available features depend on your workspace permissions.