$6,531 · $10,000 · $47,000
Surprise bills are real
Real surprise bills developers have shared publicly. By the time the invoice arrives, the money is already spent.
OpenAI, Anthropic, Gemini & OpenRouter · Free during early access
One fuse for all your AI spend
Billfuse tracks your OpenAI, Anthropic, Gemini and OpenRouter usage in one place, holds it to a single budget, and alerts you by email or Slack before a runaway agent turns into a surprise bill.
Sign in with GitHub. No card needed.
billfuse · example dashboard
This month
$412.80/ $500
83% usedprojected $571
- OpenAI$268.30
- Anthropic$144.50
80% of your monthly budget reached. Alert email sent.
On Oct 12, gpt-4o spend was 5.1× its daily average.
The problem
AI spend fails silently, then all at once
Coding agents and LLM APIs make it easy to spend money by accident. The tools that should warn you weren't built for it.
1 provider per limit
Vendor limits stop at their own border
OpenAI's spending limit only sees OpenAI. Anthropic's only sees Anthropic. Your budget covers all of them, and nothing enforces it across them.
10–12 hour lag
Billing data arrives late
Billing dashboards trail real spend by half a day. An agent stuck in a loop can burn through a month's budget before the numbers move.
Features
See it, cap it, understand it
Everything you need to keep AI spend predictable, across every provider you use.
One usage dashboard
Spend from every provider in one view, so you stop switching between billing pages.
- Month-to-date total and budget left
- Projected month-end spend
- Breakdown by provider, model and day
Budgets and alerts
Set one monthly budget across all providers, plus optional limits for each provider, and get an alert at 50%, 80% and 100% of each.
- Email and Slack alerts at 50%, 80% and 100%
- Each alert sent once per month, not on every sync
- Optional: pause OpenAI and Anthropic at 100%
Plain-English explanations
When spend jumps, Billfuse tells you which model and which day caused it.
- Your most expensive model and day
- How far a spike is from your normal
- One sentence, no spreadsheet
How it works
Three steps to set up
- 1
Connect your accounts
Paste an admin key for OpenAI or Anthropic, a management key for OpenRouter, or a Google Cloud service account key for Gemini. Keys are encrypted on our servers and used to read usage and cost data, and to pause spending only if you turn that on.
- 2
Set a budget
Pick one monthly limit that covers every provider. Billfuse warns you at 50%, 80% and 100%.
- 3
Ship without worrying
Billfuse keeps checking your usage in the background and tells you, in plain words, where the money went.
Pricing
Free while we're in early access
Every feature is free during early access. Later there will be one paid plan at $9 a month, announced on this site and by email at least 30 days before it starts.
Billfuse
Free during early access
- OpenAI, Anthropic, Gemini and OpenRouter spend in one dashboard, synced every hour
- Monthly budgets, overall and per provider, with alerts at 50%, 80% and 100%
- Alerts by email and Slack, plus an optional weekly summary
- Optional automatic pause of OpenAI and Anthropic when a budget runs out
- A plain-English explanation of what drove your spend
- Your spend in the terminal with npx billfuse
No card needed. Delete your account anytime in Settings.
FAQ
Questions, answered
Is my API key safe?
Billfuse needs an admin key for OpenAI or Anthropic, a management key for OpenRouter, or for Gemini a Google Cloud service account key that can read your billing export. These keys can do more than read usage, so we treat them like passwords. Each key is encrypted with AES-256-GCM before it is stored, is only decrypted on our servers when syncing, and is never sent back to your browser. Billfuse uses it to read your usage and cost data, and to pause and resume OpenAI and Anthropic spending only if you turn on automatic pausing or click Pause now. You can disconnect a key at any time, which deletes it.
Which providers do you support?
OpenAI, Anthropic, Google Gemini and OpenRouter. Gemini spend is read from your Google Cloud billing export in BigQuery, which takes about 10 minutes to set up. If you call other models (Google, Mistral, DeepSeek and more) through OpenRouter and pay with OpenRouter credits, their spend shows up under OpenRouter. Spend on your own provider keys inside OpenRouter (BYOK) is billed by that provider and isn't counted. OpenRouter reports each day once it's over (UTC), and Google's billing data can take up to a day, so their spend, and any alert it triggers, arrives about a day later than the others.
How much does it cost?
Nothing while Billfuse is in early access. Later there will be one paid plan at $9 a month for everything. We'll announce it on this site and by email at least 30 days before it starts, and you can delete your account anytime.
Can I stop using it?
Yes, anytime. There's nothing to cancel during early access: disconnect your keys, or delete your account and all its data in Settings.
How is this different from the spending limits OpenAI and Anthropic already have?
Each vendor's limit only covers its own platform and runs on billing data that can lag by hours. Billfuse puts one budget across all your providers, emails you at 50%, 80% and 100% of it, and explains where the money went.
Who is Billfuse for?
Solo developers and small teams of 2 to 20 people who use AI APIs and coding agents, and want to hear about a spike before the invoice does.
Don't find out from the invoice
Put a fuse on your AI spend in a few minutes. Free during early access.