Run on your own data

Drop your AWS Cost & Usage Report.
Get back a per-team breakdown.

Everything runs in your browser. Your CSV is parsed, aggregated, and classified locally — nothing transmitted, no signup, no account. CostDNA infers ownership using the same name-pattern heuristics the full version uses for its discover step.

ⓘ This is the lightweight, in-browser path — heuristic team-discovery only, no GraphSAGE training. For full behavioural attribution (including unowned-resource detection), point the Python CLI at your account: costdna scan --aws-profile prod.

Where does the CUR come from?

  1. AWS Console → Billing & Cost Management → Data Exports (formerly “Cost & Usage Reports”).
  2. Create an export with resource IDs enabled. Save to S3.
  3. Download a single month’s gzipped CSV from the bucket. Decompress.
  4. Drop the resulting .csv here.

Don’t want to set up a CUR? You can also export a small slice from Cost Explorer → Reports → Custom → CSV, but the resource-ID column will be missing on most line items. Some inference will work, most won’t.