CostPane Join the waitlist →

Know what every client's AI costs you. Bill it without a spreadsheet.

CostPane gives AI agencies one pane per client: AI usage, budgets, alerts and client-ready reports, for teams that run LLM features for many clients at once.

In early development. There's no product to log into yet. The pre-order pays for your first founding month. It's refunded on request at any time before you're onboarded, and nothing more is charged until you confirm.

Sample data for illustration

The problem

One OpenAI bill, twelve clients. A runaway loop on a Sunday. A client asking why this month cost more. You lose an afternoon a month to exports and spreadsheets, and sometimes you eat the overage.

How it works

1

Tag every call

Route model calls through a proxy, or add one client tag in your code.

2

Set budgets

Give each client a budget, alerts and an optional hard stop.

3

Report and bill

Send each client a branded monthly report and export usage to your invoices.

What you get

Who it's for

AI agencies and dev shops of 3–30 people with AI features live for several clients at once.

Planned pricing

PlanPer monthIncludes
Starter€39Up to 5 client workspaces, alerts, CSV export
Agency€79Up to 15 clients, branded reports, markup rules
Studio€159Up to 40 clients, team seats, API

The first 10 founding agencies get 50% off for 12 months and hands-on setup.

FAQ

Do you store prompts?

Only metadata by default: model, tokens, cost, client tag and timestamps.

How is this different from Langfuse or Helicone?

They track engineering usage. CostPane is built around clients, budgets and invoices.

What if my clients pay with their own API keys?

You can still import their usage for reporting.

Can I get my pre-order back?

Yes, on request at any time before you're onboarded. Just reply to the receipt email.

Tell us how you bill AI today

Five questions, two minutes. Anonymised results are shared with everyone who answers.