BurntAI

Your team's AI bill: 2 dashboards, 0 answers

BurntAI joins any AI provider on one login — GitHub Copilot and Anthropic today — so every developer's spend lands in one row, checked against their budget and forecast to month end.

burntai.dev/overview
Total spend
$10,812
+12% MoM
GitHub Copilot
$3,180
+4% MoM
Anthropic API
$5,232
+18% MoM
Anthropic seats
$2,400
+9% MoM
Sort any column. Demo data, real behaviour.
alice
$42$210$100$352112% • over
priya
$31$88$100$21973% • under
mei
$29$54$100$18361% • under
tomas
$36$0$100$13645% • under
ravi
$38$96$13444% • under
jonas
$45$71$11639% • under
Projected month-end
$14,480
13% over budget
FIG 0.1 — Every developer's Copilot spend, Claude API spend and Claude seat fee, combined and checked against budget.
  • Copilot AI credits, Claude tokens and Claude seat fees, joined on the same GitHub login
  • One row per person: every provider handle they hold, one combined cost
  • A month-end forecast that names the overspender while you can still act
  • Totals that reconcile to the GitHub and Anthropic consoles, to the cent

You rolled out AI. Now nobody can tell you what it costs, per person or in total.

AI coding tools went from a pilot budget to a real line item in one year, and finance noticed. But Copilot lives in GitHub billing, Claude Code lives in the Anthropic console, and neither one knows the other exists. So every month you hop between two dashboards, paste numbers into a spreadsheet, and still cannot answer the question finance keeps asking: what are we spending, and on whom?

Today

  • You can't even see the org's total AI spend without adding two consoles up by hand
  • The bill grows every month and you find out after it is paid, not before
  • You cannot tell a heavy user from a runaway key, so you cannot act
  • Claude subscription seats bill in a third place and attach to nobody, so the person working entirely through a seat looks free
  • You keep paying for Copilot seats that nobody has touched in weeks

With the dashboard

One screen. All providers. Every person, ranked by real spend, checked against their budget, with a forecast for where the month is heading.

  • Run the monthly AI review in minutes instead of an afternoon
  • Catch the overspender while there is still time to talk to them
  • Walk into the budget meeting with numbers that match the provider consoles

The vision

One place to see, understand, and control every dollar your organization spends on AI, no matter which provider or which person.

Version 1 reads GitHub Copilot and Anthropic (Claude Code). That is where your AI spend is today. The dashboard is built to grow past them. The goal is to read any provider your team adopts next, Gemini, OpenAI, Cursor, and whatever comes after, in the same single view. You add a tool, it lands here, and you never go back to hopping between dashboards to answer one question: what are we spending, and on whom?

FIG 0.2 — How it works

From two consoles to one row per person in four steps

01

Set it up yourself, in two minutes.

Create an account and paste your GitHub and Anthropic keys — we only ever read with them, never write back. That's it: nothing to install, no config file, no call to book.

2 minutes

02

Point each person's handles at one profile.

Say which GitHub login, Anthropic API key and Claude seat belong to the same colleague. Three partial identities collapse into one person, and stay merged.

Once, per person

03

See it all in one place.

Copilot spend, Claude API spend and Claude seat fees land in one view, per person and as a running total.

Instant

04

Set budgets, watch the forecast.

Give each person a monthly budget and catch who is trending over before month end.

Set and forget

FIG 0.3 — What's inside

Built to answer the question finance keeps asking

Every handle and every seat, on one person

One colleague can hold a Copilot seat, an Anthropic API key and a Claude subscription seat — three handles, three partial bills. All three are joined on the same GitHub login, so each person shows up once with a single combined cost, seat fee included and not just metered tokens.

aliceCopilot $42Claude API $210Claude seat $100$352 total

One person, every handle, metered spend and seat fee in one combined cost.

A month-end forecast, not just a receipt

Set a monthly budget per person. The dashboard tracks spend against it, counts who is over, and projects where each person lands by month end from their current run rate. You see the overspender before the invoice does.

Which models your money actually goes to

Spend broken down per model, across all providers, on one screen. See when the team is reaching for the most expensive model on work a cheaper one handles — and prove it with a number instead of a hunch.

Numbers finance can trust

The totals match what GitHub and Anthropic bill you, so you can put them straight in front of finance without second-guessing. When a figure is an estimate, the page says so plainly.

The monthly AI review done in minutes, not an afternoon of console-hopping
Overspend caught before the invoice, while you can still act
Finance can read it without a GitHub or Anthropic login of their own
Every dollar lands on a named person, and whatever cannot be placed is listed rather than dropped

FIG 0.4 — Why switch

Against what you do today

Not against a rival dashboard. Against the two consoles and the spreadsheet, which is what you are using right now.

The questionGitHub billing + Anthropic console + a spreadsheetBurntAI
What did the org spend on AI this month?Open two consoles, add them up by handOne number, on load
What did this one developer cost?No per-person figure exists in either console — not possible todayOne row: Copilot, Claude API and their Claude seat
Does one colleague with three handles show up once?Three partial identities, and nothing joins them — not possible todayOne person, one row, every handle merged
What is each Claude subscription seat costing us?A subscription line item attached to nobody — not possible todayThe seat fee, on the person who holds it
Are we using the right model for the job?Neither console breaks spend down per model — not possible todaySpend per model, all providers, one chart
Who is over their budget?A column you maintain in a spreadsheetTracked per person, counted for you
Where does the month land?You find out when the invoice arrives — not possible todayRun-rate forecast, per person, mid-month
Which Copilot seats are dead?Not surfaced by the billing page — not possible todayRanked by last activity
How long does the monthly review take?An afternoonMinutes

Why I built this

I run an engineering team that rolled out GitHub Copilot and Claude Code in the same quarter. AI spend was suddenly real, and no single place showed me who was spending it. Every month meant opening two consoles, exporting numbers, and still not having a per-person view I could act on.

So I built the tool I wanted. One screen that combines all providers per developer, checks each person against a budget, and forecasts month-end. It started as an internal dashboard. Enough other engineering leaders had the same problem that I opened it up.

Jiri Sifalda@jsifalda

Pricing

$39 per seat, per month

You bring your own GitHub and Anthropic access, we host and run it for you.

Team

$39/seat / month

  • Full per-person spend across Copilot, the Claude API and Claude seats
  • Budgets, remaining, and month-end forecast
  • Totals reconciled to the provider consoles
  • Spend per model, across all providers
  • Team-level rollups and priority support

FAQ

What engineering leaders ask first