detect_data_drift
Compare current vs. previous period; flags shifts >20%
How to use it
detect_data_drift is exposed by the GrowthOS MCP server. Add the server to your MCP client (Claude Desktop, Cursor, Windsurf and others), and the detect_data_drift tool becomes available to the model automatically. See the full listing for setup details and every tool this server provides.
Install GrowthOS
docker run -p 8000:8000 growth-osOther tools in GrowthOS (39)
User segmentation: Active / At-Risk / Churned
Step-by-step funnel conversion rates
Anomalies + likely cause hypotheses + investigation actions
Apply approved aliases (high-risk rules blocked unless force=true)
Campaign-to-revenue attribution bridge
Coverage gaps, unmapped keys, alias rule usage
Read-only review: coverage deltas, risk flags, audit trail
Campaign ranking by spend, efficiency, and risk
Attributed revenue and ROAS per channel
CAC + LTV + ROAS with automated invest / cut / watch classification
Cohort retention matrix (weekly or monthly) vs. B2B SaaS medians
CAC, LTV, and ROAS by channel with benchmark comparison
Schema, sample rows, and column statistics for any table
Spike/drop detection in any metric time series
Detect metric drift and optionally fire a webhook notification
Board-level executive summary
Forecast spend and daily active users for the next N days
Forecast any metric column using linear regression or exponential smoothing, with confidence bands
Side-by-side funnel conversion for two time periods with Δ (pp) column
Funnel + churn in one response with prioritized next-step recommendations
Weekly growth KPI summary with period-over-period comparison
Server status, loaded tables, DuckDB version, timestamp
Check data freshness across all loaded tables
List configured connectors and their readiness status
List all available data tables
Spend, clicks, impressions from synced Meta data
Plain-English prose narrative with headline, findings, and recommendation
Paid spend vs. Stripe revenue comparison
Preview coverage lift before applying any changes
Inspect active rules and full audit history
Undo specific alias changes by ID
Execute a read-only SQL query (DuckDB dialect); supports offset/limit pagination
Preview weekly growth report and optionally deliver it via webhook
Revenue summary from synced Stripe data
Auto-suggest high-confidence alias rules from your data
Sync Meta Ads campaigns and daily insights
Sync Stripe invoices, subscriptions, and events to local CSV
Validate tables against GrowthOS schema contracts
Full weekly growth review workflow