See exactly what's wrong
with your data stack
Connect your GitHub repo, dbt project, and cloud warehouse. Our AI scans your entire codebase and infrastructure, then sends you a comprehensive PDF report with scores, findings, and prioritised recommendations.
Read-only access · We never see your data rows · Credentials purged after analysis · Results in under 10 min
Cross-validated by multiple AI models
Your infrastructure is independently analysed by multiple AI models and their findings are merged. More coverage, fewer blind spots.
We never see your data — GDPR safe
We only read schema definitions, query patterns, and code. Your actual data rows are never accessed or transmitted. Credentials are encrypted and purged immediately after analysis.
Report in your inbox in under 10 minutes
From payment to PDF, the entire analysis typically takes 5–8 minutes. No waiting around — you'll have actionable findings before your next meeting.
Your report includes a Fintel Insight Score™ across 4 dimensions
Example scores — yours will reflect your actual infrastructure
How it works
Enter your email
No account needed. Your email is used to send your report.
Connect your data sources
Read-only OAuth for GitHub. Credential forms for your warehouse. Nothing is stored after analysis.
Pay once, securely
$99. Handled by Stripe. No subscriptions, no surprises.
Receive your report
A comprehensive PDF lands in your inbox in under 10 minutes. Scored, prioritised, actionable.
Sample findings
Every finding includes the severity, the affected file or table, a plain-English description, and a concrete fix. Here's what a typical report looks like.
Full table scan on orders table — 2.4 TB / day
daily_revenue_report.sql queries the orders table without a partition filter on `created_at`. At $6.25/TB this costs ~$5,475/year.
Fix: Add `WHERE created_at >= DATE_SUB(CURRENT_DATE, INTERVAL 90 DAY)` and ensure the table is partitioned on this column.
3 Supabase tables have RLS disabled
user_events, api_logs, and stripe_webhooks have Row Level Security disabled. Any holder of the anon key can read all rows via the REST API.
Fix: Enable RLS and add deny-all policies. Grant explicit access only to authenticated users who own the row.
17 dbt models have no tests defined
Models in marts/finance/ have no not_null, unique, or relationship tests. Schema drift or upstream NULL propagation would go undetected.
Fix: Add at minimum not_null and unique tests to primary keys, and accepted_values tests for status/type columns.
Real reports typically contain 10–35 findings across all 4 categories. Cost & FinOps findings include estimated annual savings.
What we audit
Four categories, each scored out of 10. Every finding includes the affected file or table, a plain-English description, and a concrete recommendation.
- •Full table scans and missing partition filters
- •Over-materialized dbt models
- •Expensive query patterns
- •Unused datasets and storage bloat
- •dbt test coverage gaps
- •Missing not_null / unique / relationship tests
- •Source freshness monitoring
- •Null handling and schema drift risks
- •Exposed PII columns (email, SSN, DOB, phone)
- •Overly broad IAM roles
- •Hardcoded credentials in code
- •Public tables and missing row-level security
- •Undocumented dbt models
- •Dead code and unused models
- •Naming inconsistencies
- •Circular dependencies and deeply nested CTEs
Simple, one-time pricing
No subscription. No account. Pay once, get your report. Repeat as many times as you like.
per audit, one-time
- ✓Full codebase + warehouse scan
- ✓4-category scored PDF report
- ✓Prioritised findings (Critical → Low)
- ✓Estimated cost impact per finding
- ✓Actionable recommendations
- ✓Report in your inbox in under 10 minutes
- ✓Read-only · credentials purged · GDPR safe
Fintel Insight is a product of Fintel Analytics
Found issues you need help fixing?
Get a free consultation →Your data stays private
Read-only access · credentials encrypted and purged after analysis · no data rows ever transmitted