Skip to main content

FAQ & Troubleshooting

Common questions and solutions for the Fullstory MCP server.

Getting started​

I added the server but my client can't connect. What should I check?​

Confirm the server URL is https://api.fullstory.com/mcp/fullstory with no trailing slash or typo. If everything looks correct, restart your AI client—most clients only load MCP server configurations on startup. If you haven't completed the OAuth authorization flow yet, look for a browser prompt from your client. See the Authentication page for full configuration examples. If your client reports that the protected resource does not match the expected URL, see the next question.

My client reports "Protected resource does not match expected." What should I do?​

Some MCP clients require the server URL in your configuration to match the URL the Fullstory MCP server advertises. If your org uses the EU data center and you configured https://api.fullstory.com/mcp/fullstory, the client may reject the connection when the server reports https://api.eu1.fullstory.com/mcp/fullstory. This can happen on first connect, or after a successful first connect when the client reconnects.

To fix this, set the URL to your data center:

  • EU (app.eu1.fullstory.com): https://api.eu1.fullstory.com/mcp/fullstory
  • US (app.fullstory.com): https://api.na1.fullstory.com/mcp/fullstory

You can also tell which data center you use from the prefix on your API key (eu1. or na1.). The global URL (https://api.fullstory.com/mcp/fullstory) still works for many clients—Fullstory routes you to the correct data center automatically. Use the regional URL if you hit this mismatch.

My client shows "connected" but no tools are available. What's wrong?​

The most common cause is that StoryAI features are turned off for your Fullstory org. Fullstory MCP requires StoryAI; when it's opted out, the OAuth handshake still completes successfully (so your client reports the server as "connected") but the server exposes no tools.

To fix this, an account Admin should:

  1. In Fullstory, navigate to Settings > Account Management > StoryAI Features.
  2. Confirm that StoryAI features are enabled. If they're off, turn them on.
  3. Restart your MCP client (or disconnect and reconnect the Fullstory MCP server). Tools should appear within a few seconds.

If you don't see the StoryAI Features settings page, you don't have the role required to change it—ask an Admin in your org.

If StoryAI is already enabled and you still see no tools, confirm an account admin has enabled the Model Context Protocol (MCP) toggle in Settings > Account Management > StoryAI Features.

Which AI clients are supported?​

The Fullstory MCP server works with any client that supports the Model Context Protocol. This includes Claude, Cursor, VS Code with Copilot, and others. See our client setup guides for configuration instructions specific to each client.

What's the difference between the Fullstory connector in Claude and the fullstory-skills plugin?​

Both give you the same MCP tools. The fullstory-skills plugin additionally includes the Fullstory skills (general-analysis, comparisons) and the session-context agent, which guide Claude through structured analytics workflows. The Claude connector directory listing is available in Claude (app/web) only—it doesn't apply to Claude Code, which uses the plugin. See Claude for details on both paths.

Is Fullstory MCP the right MCP for connecting AI agents to support and sales tools?​

Fullstory MCP is built for analytics—asking natural-language questions about your aggregate behavioral data. If you want to surface context about an individual user's session to an AI agent for customer support or sales use cases, use Fullstory for Customer Agents instead, which is a separate MCP integration purpose-built for that workflow.

Data & privacy​

Does Fullstory MCP respect my fine-grained access control (FGAC) setup?​

No. Orgs that require FGAC will not be admitted into the Early Access, as this will be supported in a later version.

Does the MCP respect my Fullstory privacy rules?​

Yes. Your existing masking settings, exclusion rules, and element blocking all carry over. The MCP server operates with the same privacy enforcement as the Fullstory UI and API: it does not bypass any rules configured in your account.

Usage​

Can I save segments or metrics I create through the MCP?​

To protect your configuration during Early Access, you cannot programmatically save segments or metrics. Instead, if a segment or metric is created during your session, the compute_metric tool returns a link to the unsaved object in the Fullstory Analytics app for a user to manually save. If the segment or metric is not saved during the MCP session, it will be discarded once your session ends.

Are there rate limits?​

Yes. The Fullstory MCP server is limited to 3 requests per second, with a burst allowance of 20 requests. This limit is applied per-org and per-MCP-endpoint. If you encounter throttling or receive rate-limit errors, space out your requests or reduce the scope of your queries (narrower time range, more specific segment).

MCP rate limits are separate from and independent of the public Server API rate limits—the two are tracked and enforced separately.

Can I run recurring queries automatically?​

The MCP server is a query interface: it responds to requests but does not schedule them. Automation depends on your AI client. Some clients support scheduled prompts, scripted workflows, or integration with automation tools like cron jobs or CI pipelines. As a reminder, all queries are subject to rate limits.

Early Access​

What are the Support SLAs for Fullstory MCP Beta?​

Fullstory MCP Beta is part of our Early Access program and is not covered by Fullstory's standard Support SLAs. Early Access features are opt‑in: customers can use them before general release in exchange for sharing feedback with the product team. Once a feature reaches general availability (GA), it becomes covered under Fullstory's standard Support SLAs.