FullStory MCP Server

Connect Fullstory behavioral analytics, session replays, segments, and frustration signals to AI agents through a hosted MCP server.

Analytics by Fullstory OAuth2 active

Connect (Remote Server) Docs

Overview

Fullstory MCP is a provider-hosted Model Context Protocol server that lets AI clients query Fullstory product analytics data in natural language. It bridges AI assistants like Claude, Gemini, ChatGPT, Cursor, and VS Code Copilot directly to behavioral intelligence, session replays, segments, and metrics so teams can ask questions about user experience instead of building custom queries in the Fullstory UI.

The server organizes capabilities around three workflows: frustration detection (rage clicks, dead clicks, errors, crashes, form abandonment), quantitative behavior measurement (user counts, trends, error frequency), and session analysis (retrieving session recordings with AI-generated summaries of user struggles). It also exposes administrative tools for managing users, segments, webhooks, recording settings, and event imports. The server honors existing Fullstory privacy settings, including masking rules, element blocking, and exclusion rules.

Fullstory MCP is currently in private beta and requires account enablement. It is available to existing paid Fullstory customers who join the beta program; non-customers can apply for General Availability notifications. Rate limits apply during beta, and Fine-Grained Access Control (FGAC) is not yet supported.

Tools

Tool	Description
`Get Sessions`	Retrieve session replay URLs for users by various identifier types.
`Get Session Events`	Retrieve raw, untransformed event data for a specific session.
`Generate Summary`	Produce an AI-powered summary of a Fullstory session using configured prompt profiles.
`Generate Context`	Generate AI-optimized session context from a Fullstory session for downstream reasoning.
`Get Segments`	List all user segments with IDs, names, and types.
`Create Segment Export`	Schedule a data export for a segment and return an operation ID to track progress.
`Get Users`	Retrieve user profiles by internal ID, app ID, or list all users.
`Upsert User`	Create or update a user profile with new property values.
`Batch Import Users`	Upsert up to 50 users per batch.
`Delete User`	Permanently remove a user and their associated session and event data.
`Create Event`	Create a server-side custom event for business activities like purchases or signups.
`Batch Import Events`	Bulk import up to 50 custom events per batch.
`Get Batch Status`	Monitor batch import job progress and success metrics.
`Get Operations`	Check status of async operations such as segment exports.
`Create Summary Profile`	Create a session summary prompt profile defining AI processing parameters.
`Get Summary Profiles`	List or retrieve session summary profile configurations.
`Delete Summary Profile`	Permanently remove a session summary profile.
`Get Block Rules`	Retrieve recording block rules that prevent specific elements from being captured.
`Get Domain Settings`	Access domain and subdomain recording configurations.
`Get Geo Settings`	Review geographic recording settings for compliance.
`Get Recording Features`	List enabled capture features such as console logs and network requests.
`Get Session Targeting`	Access targeting rules controlling which sessions are recorded.
`Get Webhooks`	List configured webhook endpoints with event types and status.
`Get Webhook Event Types`	Display available webhook event types for subscriptions.

Setup Guide

Prerequisites

A paid Fullstory account with MCP beta access enabled (join the waitlist at fullstory.com/platform/mcp/)
An MCP-compatible client: Claude Desktop, Claude Code, Cursor, VS Code with Copilot, or Gemini CLI

Server URL

https://api.fullstory.com/mcp/fullstory

Note the exact path with no trailing slash.

Claude Desktop / Claude.ai setup

In Claude, go to Settings → Connectors
Click Add custom connector
Enter:
- Name: Fullstory
- Server URL: https://api.fullstory.com/mcp/fullstory
Save the connector. Claude will trigger an OAuth flow to authorize with your Fullstory tenant.

Generic MCP client config

For clients that use a JSON config file with remote MCP servers:

{
  "mcpServers": {
    "fullstory": {
      "url": "https://api.fullstory.com/mcp/fullstory"
    }
  }
}

After saving the configuration, restart your MCP client. Most clients only load MCP server settings on startup. Complete the OAuth flow in the browser prompt when the client first connects.

Verify

Ask your AI client: "What Fullstory tools are available?" The agent should enumerate the connected tools.

Use Cases

Diagnose checkout or signup friction by asking the agent to identify top frustration signals (rage clicks, errors, abandonment) for a given page or segment
Pull AI-generated summaries of specific session replays to understand why a user dropped off without manually watching recordings
Quantify behavioral metrics in conversation, for example user counts, error frequency, or trend changes across releases
Manage Fullstory data hygiene: upsert users in bulk, fire custom server-side events, or honor GDPR delete requests directly from an agent
Audit recording configuration (block rules, domain settings, geo settings, webhooks) for compliance and integration reviews

Example Prompts

"What are the biggest frustrations affecting our checkout page this week?"
"How many users hit a network error on the payment page last month, and show me three example sessions."
"Summarize this Fullstory session and tell me where the user got stuck: [session URL]."
"List all segments containing 'Enterprise' and schedule an export of the largest one."
"Upsert these 20 users into Fullstory and confirm the batch status when it completes."

Pros

Official, provider-hosted server maintained by Fullstory with OAuth auth, no self-hosting required
Broad coverage across analytics (sessions, segments, events) and admin (users, webhooks, recording settings)
Honors existing Fullstory privacy controls including masking, element blocking, and exclusion rules
Works with any MCP-compatible client (Claude, Cursor, VS Code, Gemini CLI)

Limitations

Private beta only; requires an existing paid Fullstory account and program enablement
Rate limits apply during beta and standard SLAs do not yet apply
Fine-Grained Access Control (FGAC) is not supported during Early Access, and saved segments/metrics return unsaved links rather than persisting

Alternatives

PostHog MCP: open-source product analytics MCP with session replay and feature flag access
Amplitude MCP: behavioral analytics via Amplitude's APIs through community-built MCP servers
Mixpanel MCP: community MCP implementations for querying Mixpanel events and funnels