Connect the Fivetran MCP server to Unblocked to give AI agents live warehouse and business context

Code that reads from the warehouse rarely explains what the data means. Metric definitions, semantic models, and table metadata sit in the warehouse and the BI tool, so an agent writing a query or a transformation has only raw column names to go on. Fivetran’s Agent Context MCP server exposes the Context Layer you build with Fivetran — business definitions and documentation stored alongside your synced data — and Unblocked connects to it once so every agent can use it without its own Fivetran setup.

This is a third-party MCP server: Fivetran hosts and maintains it, and Unblocked connects to it as a verified MCP server. Nothing is ingested or indexed — Unblocked runs read-only queries against your Context Layer at the moment a question needs them, so answers reflect your warehouse as it is now.

Wiring Fivetran’s MCP server straight into an agent gives it a SQL endpoint. Unblocked’s Context Engine reasons across those results and your other sources — linking a metric definition in the Context Layer to the pull request that changed the model behind it and the Slack thread where the team debated how to count it — and returns one cited answer.

Once connected, any MCP-compatible coding agent — Claude Code, Cursor, Devin, GitHub Copilot, Codex — reaches Fivetran context through a single Unblocked MCP server connection.

What you get

• Business definitions, not just schemas — Agents read metric definitions, semantic models, and table metadata from your Context Layer, so a query starts from what the business means by a number rather than a guess at column names.
• One connection for every agent — Connect Fivetran to Unblocked once. Claude Code, Cursor, Codex, and GitHub Copilot reach your Context Layer through the Unblocked MCP server, with no per-agent Fivetran configuration.
• Read-only SQL, governed by your warehouse — Unblocked only offers Fivetran’s read-only query tool, and queries run under the warehouse role you grant to Agent Context MCP. Change the tool selection anytime from Settings → Data Sources.
• Context Engine reasoning — Context Layer results are reconciled with your code, pull requests, tickets, and conversations, so a disputed metric leads to the change and the decision behind it — with citations.

Prerequisites

• An Unblocked account (free to start).
• A Context Layer built with Fivetran Context Layer, and access to Agent Context MCP granted by your Fivetran account admin.
• An MCP-compatible coding agent (Claude Code, Cursor, Devin, etc.) to use the context in your editor.

Step 1: Connect Fivetran to Unblocked

1. In Unblocked, go to Settings → Data Sources → Connect another data source, and under Data & Analytics select Fivetran.
2. The Fivetran MCP server URL is filled in for you. Choose OAuth, click Connect, and sign in to Fivetran to authorize Unblocked.
3. Once Unblocked discovers the server’s tools, select the read-only tools it may use and click Save Settings.

Step 2: Install the Unblocked MCP server

On macOS or Linux, one command auto-detects and configures the supported agents on your machine:

On Windows, or when you prefer remote MCP, add the server manually using the configuration below, or follow the manual install guide .

claude mcp add -s user --transport http unblocked https://getunblocked.com/api/mcpsse

{
  "mcpServers": {
    "unblocked": {
      "url": "https://getunblocked.com/api/mcpsse"
    }
  }
}

https://getunblocked.com/api/mcpsse

Fivetran hosts and maintains this MCP server. Unblocked calls only the read-only tools you select, at the moment a question needs them:

• execute_sqlRuns a read-only SQL query against your connected destination warehouse, using the AGENTS schema to find the context available and query the relevant tables.

To give Claude Code, Cursor, GitHub Copilot, Codex, or another MCP-compatible AI coding agent access to Fivetran context, connect Fivetran to Unblocked and add the single Unblocked MCP server to your agent.

Unblocked acts as the MCP context layer for your Fivetran data, searching Context Layer and read-only SQL alongside your other connected sources and returning cited results through these tools:

• context_researchResearches a question across your connected sources and queries your Fivetran Context Layer when metric definitions, semantic models, or warehouse data are relevant, returning a synthesized answer with citations.
• context_get_urlsFetches the full content behind URLs your agent encounters across connected sources — Jira, Linear, Notion, Confluence, Slack, GitHub, and more.

Questions you (or your agent) can ask Unblocked about your Fivetran data:

• How is active customers defined in our Context Layer, and which pull request last changed the model behind it?
• Which warehouse tables describe subscription revenue, and is there a ticket about the discrepancy finance reported?
• Does the metric this dashboard query uses match the definition the data team agreed on in Slack?
• What does the Context Layer say about the orders table before I change this transformation?