Connect the Amplitude MCP server to Unblocked to give AI agents live product analytics context

When a funnel drops or an event stops firing, the answer is split between Amplitude and the codebase: the chart shows the change, but the tracking call, the feature flag, and the release that caused it live somewhere else. Amplitude’s MCP server exposes charts, the tracking plan, experiments, and flags to AI agents, and Unblocked connects to it once so every agent can ask about product behaviour without its own Amplitude setup.

This is a third-party MCP server: Amplitude hosts and maintains it, and Unblocked connects to it as a verified MCP server. Nothing is ingested or indexed — Unblocked calls the read-only tools you select at the moment a question needs them, so answers reflect your Amplitude projects and their events as they are now.

Wiring Amplitude’s MCP server straight into an agent gives it chart queries. Unblocked’s Context Engine reasons across those results and your other sources — following a drop in a conversion chart to the event the instrumentation stopped sending, the pull request that renamed it, and the Slack thread where the change was agreed — and returns one cited answer.

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

What you get

• Live charts and ad-hoc queries — Agents read saved charts and run event segmentation, funnel, and retention queries against your Amplitude projects as they are right now, not a stale export.
• Tracking plan next to the code — Event and property definitions from your Amplitude taxonomy sit alongside the code that sends them, so agents can check instrumentation before they change it.
• Read-only, tool-level control — Only read-only Amplitude tools are offered — nothing that edits charts, cohorts, flags, or the tracking plan — and you choose which of them Unblocked may call. Change the selection anytime from Settings → Data Sources.
• Context Engine reasoning — Amplitude results are reconciled with your code, pull requests, tickets, and conversations, so a shift in a metric leads to the release or experiment behind it — with citations.

Prerequisites

• An Unblocked account (free to start).
• An Amplitude account with access to the projects you want to query, MCP access allowed for your organization, and a role that grants the Use MCP (read) permission.
• An MCP-compatible coding agent (Claude Code, Cursor, Devin, etc.) to use the context in your editor.

Step 1: Connect Amplitude to Unblocked

1. In Unblocked, go to Settings → Data Sources → Connect another data source, and under Data & Analytics select Amplitude.
2. The Amplitude MCP server URL is filled in for you. Choose OAuth, click Connect, and sign in to Amplitude 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

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

• searchSearches for charts, dashboards, notebooks, experiments, events, properties, cohorts, and other Amplitude content.
• query_amplitude_dataRuns ad-hoc analytics queries for event segmentation, funnels, retention, and more, or queries an existing chart by ID.
• get_amplitude_chartsReturns chart definitions and data by ID, including events, properties, and configuration.
• get_amp_taxonomyReads a project’s tracking plan — its events and properties, the changes on a branch, and activity history.
• get_experimentsReturns experiments by ID with their state, variant configuration, and decision details.
• get_flagsReturns feature flag definitions by ID, including variants and configuration.

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

Unblocked acts as the MCP context layer for your Amplitude data, searching charts, dashboards, and product analytics alongside your other connected sources and returning cited results through these tools:

• context_researchResearches a question across your connected sources and calls the Amplitude tools you selected when charts, events, or experiments 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 Amplitude data:

• Why did the checkout funnel in Amplitude drop after Tuesday’s release, and which pull request touched the checkout events?
• Is the signup_completed event in our tracking plan still sent anywhere in the codebase, and who last changed it?
• What did the onboarding experiment decide, and is there a ticket to clean up its feature flag?
• Which properties does the search_performed event carry in Amplitude, and do they match what the web client sends?