Step 1
Paste trace exports
Bring JSON or NDJSON from your provider, gateway, or app logs with explicit cost or supported model token usage.
Free AI spend attribution audit
AI Cost Attribution Audit
Break down AI spend by team, project, and customer.
Built for AI platform leads and FinOps owners who need the bill mapped to the right owner before invoice review.
Paste JSON or NDJSON traces from provider, gateway, or app logs to see who drove the bill, export the current breakdown, and hand finance a clean table.
Saved reports? Join the waitlist.
What this audit gives you
See who drove AI spend before finance asks
- Paste JSON or NDJSON traces from provider, gateway, or app logs.
- Group positive AI cost by team, project, or customer.
- Export the current breakdown as JSON or CSV.
Attribution breakdown
Grouped by team · last 7 days
- Total cost
- $526.50
- Grouped by
- Team
- Matching
- 3 teams
| Team | Cost | Input tokens | Output tokens | Requests |
|---|---|---|---|---|
| ml-eng | $312.80 | 412K | 198K | 487 |
| finops | $124.50 | 168K | 77K | 312 |
| growth | $89.20 | 121K | 57K | 224 |
Live attribution preview
Run the free attribution check
Paste usage traces, switch between team, project, and customer ownership, and export the current view as soon as the breakdown answers the bill.
Awaiting trace input.
Example: [{"team":"ml-platform","project":"copilot-search","customer":"acme-co","model":"gpt-4.1","inputTokens":12000,"outputTokens":1400}]. Results update live as you paste or type.
Live Preview
Live preview
Add valid input to populate this result panel.
Paste JSON or NDJSON traces from your provider, gateway, or app logs to see AI spend grouped by team, project, or customer.
Why this tool
Attribution wedge
Why AI spend ownership gets blocked
Shared AI infrastructure makes invoices ambiguous. Provider charges arrive as one total even when the underlying requests came from different teams, separate product projects, and customer-facing workloads with different business owners.
The fastest fix is not another generic observability dashboard. It is a narrow attribution review that turns raw traces into a clear owner table before finance, platform, and engineering start arguing over the same bill.
- Invoices collapse multiple teams, projects, and customers into one AI budget line with no obvious owner.
- Provider traces often already contain the cost detail you need, but it stays buried in raw logs until someone normalizes it.
- Attribution turns a billing surprise into a concrete review: who used the spend, which project drove it, and which customer workload needs follow-up.
Three-step flow
How the attribution audit works
Step 2
Switch the grouping
Recompute the same spend by team, project, or customer without reloading the page or rewriting the input.
Step 3
Export the current breakdown
Copy JSON, download JSON, copy CSV, or download CSV once the attribution table is ready to share.
What this route answers
What you can resolve before the next invoice review
Team ownership before finance review
Show which internal team actually consumed the spend instead of escalating a single blended AI invoice.
Project-level chargeback split
Separate one product initiative from another before deciding where the next optimization or chargeback lands.
Customer workload visibility
Show whether customer-facing AI features, pilots, or premium workloads are driving the bill.
Reusable export for saved reviews
Move the current breakdown into recurring reviews once you need a saved monthly record.
FAQ
Questions teams ask before trusting the breakdown
Route out
Choose the next attribution path
Stay on the free audit for a one-off spend question, then move into pricing, saved reports, or the broader cost auditor when the same question keeps coming back.
Saved attribution reports
Move from one-off audits to saved attribution reviews
Use the public audit to map spend once, then move into saved reports or pricing when the same review needs to happen every month.