Usermaven MCP server | AI agent analytics

Q: What is an MCP server?

An MCP server uses Model Context Protocol (MCP), an open standard that allows AI tools to connect with external apps and data. Usermaven's MCP server lets compatible AI tools access your Usermaven data so you can analyze performance and take approved actions directly through AI.

Q: What can I do with Usermaven MCP?

As an AI-powered marketing attribution platform, Usermaven lets you use MCP to ask AI questions about attribution, revenue, CAC, LTV, and campaign performance, along with product data, funnels, retention, and customer journeys. You can also create or update dashboards, funnels, segments, journeys, reports, conversion goals, and attribution views with your approval.

Q: Which AI tools work with Usermaven MCP?

Usermaven MCP works with ChatGPT, Claude, Cursor, Codex, and other AI clients that support remote MCP servers. Once connected, these tools can access the Usermaven workspaces and data you authorize.

Q: Can I connect Usermaven to ChatGPT with MCP?

Yes. Add the Usermaven remote MCP server to ChatGPT and complete the Usermaven authorization process. Once connected, you can ask ChatGPT questions about your analytics and use the Usermaven capabilities you've permitted.

Q: What data can AI agents access through Usermaven MCP?

AI agents can access the Usermaven data you authorize, including website and product analytics, marketing attribution, CAC, LTV, ARR, revenue, conversions, funnels, customer journeys, retention, segments, dashboards, reports, and event taxonomy. Access is limited to the workspaces and permissions you select when connecting MCP.

Q: Can Usermaven MCP make changes to my analytics?

Yes. If you enable write permissions, your AI can create or update funnels, segments, journeys, retention and trend reports, dashboards, conversion goals, and attribution views. Write actions only happen when you explicitly request them and approve the action.

Q: Is Usermaven MCP secure?

Usermaven MCP gives you control over which workspaces and permissions an AI connection can access. Read and write permissions are managed separately, and workspace admins can control whether MCP and write access are allowed. Write actions also require your explicit approval before any changes are made.

Q: Can I use Usermaven MCP across multiple workspaces?

Yes. You can authorize multiple Usermaven workspaces on a single MCP connection and switch between them through your AI assistant. This is especially useful for agencies and teams managing multiple brands, products, or client workspaces.

Q: Do I need technical knowledge to set up Usermaven MCP?

No. Add Usermaven's remote MCP server to a supported AI client, sign in to Usermaven, choose the workspaces and permissions you want to allow, and approve the connection. You don't need to manually manage API keys or build a custom integration.

Q: Does Usermaven MCP support both read and write actions?

Yes. Read actions let AI analyze your Usermaven data without changing anything. Write actions let AI create or update supported analytics, such as funnels, dashboards, segments, journeys, and reports, and always require your approval before they run.

Trusted by marketing, product, and growth teams worldwide

4.8

4.9

4.8

4.9

Powerful MCP capabilities

Bring in your data. Get answers. Take action with AI.

Connect your AI tools to Usermaven through MCP to analyze key metrics, investigate performance, and create analytics directly from your conversations.

Business metrics

Analyze CAC, LTV, ARR, and revenue with AI

Ask questions about the metrics that matter to your business and get answers grounded in your Usermaven data. Compare performance over time and investigate changes without manually pulling together reports.

Analyze CAC, LTV, ARR, and revenue
Compare key metrics across different time periods
Investigate changes and performance trends

Business metricsQ2 vs Q1

“How did CAC and LTV move last quarter?”

CACBlended, last 90 days

$214↓ 12.4%

LTVPer paying customer

$3,180↑ 8.1%

ARRAnnual recurring revenue

$1.42M↑ 22.6%

RevenueNew MRR this quarter

$96.4K↑ 4.3%

Answered from your Usermaven workspace · 4 metrics compared

Marketing attribution

Find the channels and campaigns driving real growth

Ask your AI to analyze attribution across the customer journey and uncover which channels and campaigns drive signups, paying customers, revenue, and stronger ROAS.

Compare performance across channels and campaigns
Identify sources driving your best customers
Analyze conversions and revenue by attribution model

Channel attributionLast 90 days

ATTRIBUTION MODELS

First touch ✓Last touch ✓Linear ✓

Conversions by channel across attribution models

First touchLast touchLinear

Organic search

$74.9K

Paid search

$83.7K

Referral

$28.4K

$19.2K

Organic social

$12.6K

Paid search wins on last touch, organic leads on first touch

Funnels & customer journeys

Find where users convert, drop off, and why

Explore funnels and customer journeys through natural-language questions. Understand the paths users take before converting, identify friction points, and compare behavior across different audiences.

Identify high-impact funnel drop-offs
Analyze paths leading to key conversions
Compare journeys across users and segments

Activation funnelStep 3 flagged

Signed upEntry point

12,847100% of users

↓ 44.6% dropped off5,729 users lost

Completed setupEvent triggered

7,11855.4% of users

↓ 67.1% dropped off4,777 users lost

Invited a teammateEvent triggered

2,34118.2% of users

Highest drop-off

↓ 36.5% dropped off855 users lost

Upgraded to paidConversion goal

1,48611.6% of users

TOP CONVERTING PATH

Organic search›Pricing page›Docs ×3›Upgraded

AI-powered analytics creation

Turn a conversation into the analysis you need

Go beyond querying data. Ask your AI to create or update analytics in Usermaven, then review and approve the action before any changes are made.

Create funnels, journeys, and retention reports
Build dashboards and add dashboard tiles
Create segments, conversion goals, and attribution views

Claude

MCP

Create an onboarding funnel from signup to a paid upgrade.

usermaven · create_funnelNeeds approval

I'll create a 4-step funnel called Onboarding to paid. Review the steps before I make any changes.

Signed upuser_signed_up

Completed setupsetup_completed

Invited a teammateteammate_invited

Upgraded to paidsubscription_started

Approve & createCancel

Nothing is created in Usermaven until you approve it

Multi-workspace analysis

Analyze multiple workspaces from one AI connection

Connect the Usermaven workspaces you need and switch between them directly from your AI conversation. Compare performance across brands, products, or client workspaces without reconnecting each time.

Switch between authorized workspaces instantly
Run the same analysis across different workspaces
Keep workspace access controlled by permissions

ChatGPT

3 workspaces

Acme SaaSAdminNorthwindAnalystClient · VertexViewer

Compare signups this month across my workspaces.

Acme SaaS4,182↑ 18.4%

Top source · Organic search

Northwind2,760↓ 6.2%

Top source · Paid search

Acme SaaS leads on volume, but Northwind converts paid traffic at a higher rate.

Each workspace stays scoped to your existing permissions

How Usermaven MCP works

Bring Usermaven directly into your AI workflows

Built on the open MCP standard

Usermaven provides a remote MCP server built on Model Context Protocol (MCP), the open standard for connecting AI tools with external data and capabilities. Connect once and use your Usermaven data with MCP-compatible AI clients without building a separate integration for each one.

usermaven-mcpRemote server · JSON-RPC

MCP

→initialize { protocolVersion: "2025-06-18" }

←serverInfo { name: "usermaven-mcp" }

→tools/list

←[ "query_metrics", "get_funnel", … ]

Open standard — any MCP client speaks it

Give AI access to your key metrics

Your AI can query Usermaven data directly to analyze CAC, LTV, ARR, revenue, attribution, conversions, funnels, customer journeys, retention, and more. Instead of copying numbers or exporting reports into an AI chat, you can ask questions and get answers based on your Usermaven data.

CALLusermaven.query_metrics

RESPONSELive

CAC

$214

LTV

$3,180

ARR

$1.42M

Revenue

$96.4K

Use read and write actions with AI

Usermaven MCP supports both read and write actions. Along with analyzing your data, you can ask AI to create or update funnels, segments, journeys, retention reports, trend reports, dashboards, conversion goals, and attribution views. Every write action requires your approval before any change is made.

READquery_metricsget_attributionget_funnelget_journey

WRITEcreate_funnelcreate_segmentsave_dashboardApproval required

Control access with MCP permissions

Choose exactly which Usermaven workspaces and permissions you want to make available when authorizing an MCP connection. Read and write permissions are controlled separately, workspace admins can manage MCP access, and changes cannot be made without your explicit approval.

Connection scopesSet at authorization

Active

WorkspaceAcme SaaS

ReadMetrics, funnels, journeys

WriteRequires approval

BillingNot shared

Work across MCP-compatible AI tools

Connect Usermaven to ChatGPT, Claude, Cursor, Codex, and other AI clients that support remote MCP servers. One standard connection brings your Usermaven context into different AI workflows without requiring a custom integration for every tool.

Claude

ChatGPT

Cursor

Codex

Any MCP client

MCP

One connection, every compatible client

Set up MCP connection in minutes

Add Usermaven's remote MCP server, sign in, select your workspaces and permissions, and approve the connection. Follow the Usermaven MCP setup guide for detailed instructions.

MCP SERVER CONFIG

{ "mcpServers": {

"usermaven": {

"url": "mcp.usermaven.com"

} } }

1Add server→2Sign in→3Approve scopes

FAQs