Skip to main content

​
1. Install the SDK

npm install @vgpprasad91/churnkit-sdk

​
2. Create a client

import ChurnKit from '@vgpprasad91/churnkit-sdk'

const churn = new ChurnKit({ apiKey: process.env.CHURNKIT_API_KEY! })

​
3. Identify your users

Call identify() when a user signs up or logs in. This attaches traits that the scoring model uses. 
await churn.identify('user_123', {
  plan: 'pro',
  mrr: 4900,          // monthly recurring revenue in cents
  email: 'alice@acme.com',
  signedUpAt: '2024-01-15T00:00:00Z',
})

​
4. Track product events

Track events that signal engagement (or disengagement): 
// User is active — positive signal
await churn.event('user_123', 'feature_used', { feature: 'data_export' })
await churn.event('user_123', 'dashboard_viewed')

// Potential churn signals
await churn.event('user_123', 'support_ticket_opened', { severity: 'high' })
await churn.event('user_123', 'export_all_data')

​
5. Get the risk score

const risk = await churn.score('user_123')

console.log(risk.score)          // 0.73
console.log(risk.tier)           // 'high'
console.log(risk.signals)        // ['no_login_7d', 'support_spike']
console.log(risk.recommendation) // 'Reach out to this user immediately.'

​
6. Set up a webhook alert (optional)

Get notified automatically when a user’s risk crosses a threshold: 
await churn.watch('user_123', {
  threshold: 0.7,
  webhook: 'https://yourapp.com/hooks/churn-alert',
  cooldown: '24h',  // don't fire more than once every 24 hours
})

​
What’s next?

Configuration

Timeouts, retries, observability hooks, and test mode.

Event Batcher

High-throughput event ingestion with automatic batching.

At-Risk Users

Query all users above a risk threshold for bulk campaigns.

Webhooks

Verify webhook signatures and handle payload events.