New@churnkey/mcp

Churnkey MCP Server: Bring Churnkey to your AI agent

Connect an AI agent to Churnkey with one sign-in, and it can read your retention data, build Cancel Flows and dunning campaigns, run experiments, and publish changes live—all scoped to exactly what you can already do.

MCP is the open standard for connecting AI agents to your data and tools.

View Docs
Churnkey MCPconnected
How's our Payment Recovery doing, broken down by decline reason?
Thought for 2.0s
Usedchurnkey.aggregate_payment_recoveries

Across 447 failed payments, Payment Recovery clawed back $173,419 of $214,756 at risk—an 81% recovery rate. By decline reason:

Decline reasonAttemptsAt riskRecoveredRate
insufficient_funds171$79,983$68,17585%
do_not_honor54$33,393$31,26994%
generic_decline34$21,028$18,61989%
transaction_not_allowed31$16,575$14,13485%

insufficient_funds is the volume game—38% of attempts, recovered at 85% by retrying around payday. do_not_honor quietly recovers best at 94%.

Add a final SMS nudge on payday to the insufficient_funds sequence, then publish it.
Thought for 1.5s
Usedchurnkey.add_recovery_email

Added an SMS step to the recovery draft, timed to land on payday:

insufficient_funds recovery Published
  • Email · day 1Card declined
  • Email · day 3Retry reminder
  • SMS · paydayFinal nudge

Live · pending sends rebuilt on in-flight sequences·you · mcp-oauth

Usedchurnkey.publish_recovery_blueprint

Live now. The new step sends about a day after the last, and in-flight sequences pick it up on their next send.

Send a follow-up…
Ask AI agent
try asking about:

Works with

Claude logo ClaudeCursor logo CursorChatGPT logo ChatGPTPerplexity logo PerplexityCodex logo CodexVS Code logo VS Code and more

Your retention stack lives in dashboards. Your team would rather not.

Every retention task turns into a context switch. Open the dashboard. Pick the right view. Filter by plan, by cohort, by month. Find the flow. Edit the step. Publish. Repeat.

By the time it's done, the moment for the next thing is gone. So the work piles up—or never happens.

The Churnkey MCP MCP (Model Context Protocol) is an open standard that lets your AI agent connect to tools and data sources like Churnkey. server gives AI agents permission-aware access to read, build, and publish across your entire retention stack—Cancel Flows, dunning, experiments, and analytics.

How it works

Retention work, where your team already works.

The MCP server is a tiny package that connects your Churnkey workspace to any AI agent that speaks the Model Context Protocol. Sign in once, and every question—and every change—runs against your real data, from inside the tools your team already trusts.

No new dashboard to learn. No new tab to forget about. Just retention context, on the same surface as the work.

Retention work, where your team already works.

Turn it on.

A workspace admin flips on MCP Access for your user in Churnkey → Team. It's off by default, so nothing connects until someone opts in.

Churnkey → Team → MCP Access

Sign in once.

One OAuth sign-in in your browser. You approve exactly what the agent can do on a consent screen—no API key to copy, no secret in your config.

npx @churnkey/mcp auth login

Ask in plain language.

"What's my save rate this quarter?" "Which offer is winning?" "Add a pause step to the flow." Whatever you'd ask a teammate, ask your agent instead.

"Why did Enterprise churn spike in February?"

Read the answer—or make the change.

Take the number and move on, or tell it to act: spin up an A/B test, tighten dunning, publish a flow. It works within your permissions, and logs every change under your name.

Draft updated · published · logged to audit trail

The tool catalog

One connection. The whole retention stack.

The MCP server hands your agent 54 operations across six areas of Churnkey. Each one either reads your data, builds a reversible draft, or publishes to live. Filter to see what fits.

13 tools

Cancel Flows

The cancel experience your subscribers see—its steps, offers, surveys, and audience routing.

  • Read every flow, its live/draft state, and step-by-step config
  • Rewrite copy, offers, and survey choices in a safe draft
  • Publish the draft live, or re-prioritize segments
7 tools

Experiments

Two-arm A/B tests that split a flow on live traffic and crown a winner.

  • Read per-arm results with statistical significance
  • Clone a flow into a variant and set the test up
  • Start, pause, or commit the winner to all traffic
18 tools

Payment Recovery

Dunning—the emails and texts that recover failed payments, and the revenue they reclaim.

  • Read recovered vs. lost revenue, decline reasons, and per-message engagement
  • Edit email + SMS sequences, offers, and audience rules
  • Publish campaigns, toggle them, or stop a running sequence
4 tools

Analytics & Context

Every flow interaction and its outcome—save rate, boosted revenue, and workspace context.

  • Query sessions and roll them up by any dimension
  • Pull save rate, boosted revenue, and feedback themes
  • Orient on which workspace, mode, and permissions you have
9 tools

Settings & Domains

How cancels and pauses execute in Stripe, Adaptive Offers strategy, and hosted-page domains.

  • Read billing behavior, Adaptive Offers config, and DNS state
  • Change billing settings and tune Adaptive Offers
  • Set subdomains and register custom domains
3 tools

Governance & Privacy

The attributed trail of every change an agent or human made, plus GDPR access and erasure.

  • Read the audit log of every agent and human action
  • Fulfill a GDPR data-access request by email
  • Erase a customer's data—destructive, confirm-gated
Read — looks, never changesBuild — edits a reversible draftPublish — acts on live data

Built for everyone

Churn isn't one team's problem. Neither is the work.

RevOps, founders, retention, engineers—everyone touches churn. Pick a role to see what they ask, and what the MCP server does about it.

Cohort analysis

"Break down save rate by plan for last quarter."

Sessions rolled up by plan, segment, or month—no SQL, no export, no waiting on the data team.

Dunning performance

"How's Payment Recovery doing by decline reason?"

Recovered vs. at-risk dollars, split by decline reason, card brand, or currency, with the trend baked in.

Churn drivers

"Which cancel reason grew the most last month?"

Rank the fastest-rising reasons from your Cancel Flow surveys, so you know what to fix next.

Version compare

"Did the new flow beat the one it replaced?"

Flow metrics compare two published versions side by side—save rate and boosted revenue.

Connect in a minute

Two ways to plug in.

The same server works everywhere MCP does. Point a hosted client at one URL, or run it locally with a single command—either way you sign in with OAuth, not an API key.

First, get access. A workspace admin turns on MCP Access for your user in Churnkey → Team. It's off by default for everyone.

Connect by URL

For hosted clients like Claude and ChatGPT. Add one connector URL—it signs you in with OAuth in your browser. No install, no key.

https://mcp.churnkey.co/mcp

In your client: Settings → Connectors → Add custom connector, paste the URL, and approve access.

Run it locally

For Claude Code, Cursor, and Claude Desktop. Sign in once, then register the server—it spawns on demand with your stored session.

# sign in (opens the consent screen)
npx @churnkey/mcp auth login
# then register with your client
claude mcp add churnkey -- npx -y @churnkey/mcp

Prefer config files? Drop the same block into .mcp.json or your client's config. See the docs.

Either way, you approve exactly what the agent can do on a consent screen—and every action is recorded under your name.

You're in control

Powerful, never off the leash.

An agent that can publish live changes needs real guardrails. Every one is built in.

Per-user OAuth 2.1.

Sign in as you over OAuth 2.1 with PKCE—no API keys in your config, no shared secret to leak.

Role-scoped.

An agent can never do more than you can in the dashboard. Scopes are capped to your workspace role.

You approve every scope.

A consent screen shows exactly what it asks for. PII and live-change scopes are flagged—uncheck anything before you approve.

Fully audited.

Every call is attributed to you and tagged mcp-oauth in the audit log, with before/after values.

Admin-gated.

MCP access is off by default. An admin enables it per user in Team—and can revoke any session instantly.

Drafts, confirms, and limits.

Reads are always safe. Edits land in drafts, going live takes an explicit confirm, and writes are rate-limited.

Open source and MIT-licensed—published on npm as @churnkey/mcp. Audit the code yourself.

Compound effect

Retention work, on the same surface as the work.

When the answer and the fix are both a sentence away, the loop gets tighter. Teams that bring retention into their AI workflow notice patterns earlier, ship changes faster, and compound small wins.

The MCP server is a small package. The behavior change it unlocks is not.

Why did churn spike last week?
A/B test a save offer against the current one.
Tighten the dunning cadence and publish it.
Asked, done, and shipped. Zero dashboards.

Still unsure?

Frequently asked questions

Enterprise-grade security, built in.

Churnkey meets top industry standards for data security and compliance.

Bring Churnkey to your AI agent.

The MCP server is free, open source, and ready the moment you are. Turn it on, sign in, and ask your first question in a minute.