Burnline
Burnline reads your AI agent and API usage logs on your own machine and shows what they cost, by model and by day, against a monthly budget. Prompts stay on your machine.
It is not for sale yet. A CLI and an HTML report exist as a prototype; there is no checkout and no cloud login. If you run multi-model coding agents and want launch notes, join the waitlist. Please don't send money for Burnline today.
The button opens an email to tolaria@naver.com with the subject Burnline waitlist.
Add one paragraph on how you use AI today. Checkout (Stripe or a form) will only be added once there is real demand.
How it works
Import logs locally
Point Burnline at Claude Code or JSONL usage logs. You don't have to upload prompts to get the cost numbers.
Price each call
An offline model catalog with documented rates turns tokens into dollars. The totals are meant to be reconciled against your real bill.
Set a monthly cap
Set a budget in USD. Burnline warns at 80% and flags at 100%. Export the data as CSV for your records.
Planned pricing
These are the intended paid tiers for later. None of them is live and there is no buy button.
| Plan | Price | Includes | Status |
|---|---|---|---|
| Free | $0 | 7-day window, one source, core charts | Planned |
| Pro | $19 / mo | Unlimited history, budgets, CSV, catalog updates | Planned |
| Team | $49 / mo | Shared aggregates for small teams. Not in v0.1. | Later |
What it is not
Burnline is not a full prompt observability cloud, and v1 has no enterprise SSO. If you need Helicone-scale pipelines, use those. Burnline is for builders who want accurate cost numbers with little setup.
Need help with a production mobile app now? See the mobile retainer.