AI cost control for small teams

Know which AI feature is burning money.

BurnCap ingests your LLM usage — an OpenAI or Anthropic import, a CSV, or a five-minute SDK — and shows cost per feature, customer and model. It forecasts month-end spend, alerts on runaway agents, and caps budgets before they destroy your margin.

30 days of StarterNo card · no checkout · no automatic renewal

Then Free unless you choose to upgrade · prompts never stored

burncap.app/dashboard
How it works

From raw usage to a verdict, in minutes.

01

Send usage

Pull daily usage from the OpenAI or Anthropic admin APIs, upload a CSV, or track per-request with the TypeScript SDK. Idempotent by request id — retries never double-count.

02

See the burn

Spend by day, model, feature, customer and environment. Forecasted month-end. Estimated vs provider-billed, always labeled — no fake precision.

03

Cap it

Budgets per workspace, feature, customer or environment with 50/80/100% alerts. Hard caps return budget_exceeded through the SDK — your app stays in control.

Guardrails, not surprises

One check before the expensive call.

Fails openA monitoring outage never takes down your app.
Never a proxyBurnCap never sits between you and your model provider.
Your app decidesCaps are answers, not interceptions. Enforcement stays in your code.
guard.ts
const gate = await burncap.checkBudget({
feature: "agent-runs",
customerId: user.id,
});
if (!gate.allowed) {
return res.status(402).json({ error: "budget_exceeded" });
}
// run the model call, then:
burncap.trackUsageInBackground({
request_id: completion.id,
provider: "openai", model: completion.model,
input_tokens: usage.prompt_tokens,
output_tokens: usage.completion_tokens,
});
For your AI agent

Let your coding agent wire it up.

You already build with Claude Code or Cursor. We ship dedicated prompts that hand your agent the exact SDK setup and per-provider token math for your stack — paste one in and BurnCap is integrated. Zero docs to read.

Next.js + Vercel AI SDK

Current totalUsage details, including cache reads, writes, and multi-step calls.

Node / Next.js + OpenAI SDK

OpenAI Responses or Chat Completions, with cache and reasoning separated.

Node + Anthropic SDK

@anthropic-ai/sdk — cache reads, 5m/1h writes, region, and web search.

Node + Google Gemini SDK

@google/genai — resolved model, modalities, cache, tier, and thinking tokens.

Python (FastAPI + OpenAI / Anthropic)

No SDK needed — thin httpx calls, fail-open guardrail.

What you get

Everything the invoice hides.

01

Feature & customer breakdowns

Tag events with feature and customer_id; see exactly who and what burns money.

02

Forecast & spike alerts

Month-end forecast from your trailing average. Spike, runaway-loop, expensive-model and dev-spend detectors.

03

Burn Health Score

One 0–100 readiness number for whether you'd actually catch a surprise — then a checklist of what to fix first.

04

Budgets with real teeth

Alert-only by default. Soft and hard caps exposed to your app through one endpoint.

05

Unit economics

Cost per request, per active user, per paying customer — with margin % against the revenue you set.

06

Burn report digests

A weekly or daily email recap — spend, forecast, budget status, top movers — so nobody has to remember to look.

07

Honest numbers

Every figure is labeled estimated, provider-billed or forecasted. Pricing overrides for negotiated rates.

08

Team workspaces

Invite your team, scope an API key per project, and keep every app's spend under one roof.

09

Private by default

Token counts and identifiers only. No prompts, no completions, ever. Delete all usage data anytime.

Why BurnCap

Not another observability dashboard.

Most tools log what your AI did. BurnCap is built around the money question — and for the size of team that doesn't have a FinOps department to answer it.

Budget-first, not trace-first

Observability tools are built to debug traces. BurnCap answers the question that actually keeps you up: am I going to lose money this month?

Built for teams of 1–10

Not an enterprise FinOps suite that needs a dedicated team to run. Connect in minutes — there's no setup project to budget for.

Per-customer profitability

Set the revenue you charge each customer and see real margin — the answer to "is this account worth keeping?" without a spreadsheet.

Honest, labeled estimates

Stale price tables quietly miscount by 20–40%. Every BurnCap figure is labeled and reconciled against what your provider actually billed.

Pricing

Cheaper than one surprise bill.

Included after eligible signup

30 days of Starter, already active.

One grant per eligible account. No card, checkout, or automatic renewal. Choose a paid plan only if you want to; otherwise the workspace returns to Free.

Free
$0/mo

See your AI costs for the first time.

  • 10k usage records/month · 7-day history
  • 1 project · 1 seat
  • CSV + provider imports (manual sync)
  • Budgets, caps & policy-action API
  • Spike & runaway-agent alerts (email)
  • Burn readiness score & forecast
Create free account
30 days included
Starter
$19/mo

Automated monitoring & alerts.

  • 100k usage records/month · 90-day history
  • 3 projects · 3 seats
  • Daily auto-sync for provider imports
  • Slack alerts + 5 email recipients
  • Weekly burn report (spend, budgets, movers)
  • Cost insights & what-if savings
Start 30 days free
Growth
$39/mo

Margin protection & business reporting.

  • 1M usage records/month · 365-day history
  • 5 projects · 10 seats
  • Everything in Starter
  • Margin alerts when a customer turns unprofitable
  • Daily burn report
  • Unit economics across a full year of history
Start with Starter free
FAQ

Everything you'd ask before connecting.

Do I have to route my traffic through BurnCap?

No. BurnCap is never a proxy — your model calls never pass through it. You send usage out-of-band through an OpenAI or Anthropic import, a CSV upload, or our SDK. Budget checks default to fail open if BurnCap is unreachable; fail-closed behavior is an explicit SDK choice.

How do I get my usage into BurnCap?

Three ways, and you can mix them: pull daily totals from the OpenAI or Anthropic admin APIs, upload a usage CSV, or track per-request with the @burncap/sdk. Ingestion is idempotent by request id, so retries and re-uploads never double-count.

Will BurnCap see my prompts or my customers' data?

No. BurnCap stores token counts, costs, and only the identifiers you choose to send — feature, customer_id, environment, request id. It never receives or stores your prompts or model completions. Ever.

Can BurnCap actually stop a runaway agent or a budget overrun?

BurnCap returns fresh budget state and an advisory action; your code decides what to do — block, warn, or drop to a cheaper model. BurnCap never blocks traffic itself, and the SDK defaults to fail open, so enforcement and outage behavior stay under your control.

How does the 30-day Starter access work?

Each eligible account can receive one 30-day Starter grant after signup and, when required, email verification. No card or checkout is needed, nothing renews automatically, and no charge occurs at expiry. Unless you voluntarily upgrade, the workspace returns to Free limits with its existing data intact. Existing paid or previous promotion recipients are not eligible.

Is there a free plan?

Yes — free forever, no credit card. The Free plan includes 10k ingested usage records per month, 7-day history, one project and one seat, budgets and caps, and spike and runaway-agent alerts by email.

How is billing handled, and can I cancel anytime?

Paid plans are billed monthly through Lemon Squeezy, our merchant of record. Cancel anytime from the in-app billing portal — you keep paid access until the end of your period, then revert to Free with your data intact.

See all questions
Start now

Catch the spike before the invoice does.

30 days of Starter, no card and no automatic renewal. Afterward, stay on Free unless you choose to upgrade.