Hosted version · join the waitlist
Track your OpenAI API spend, day by day
Tallypane reads your organization’s costs from OpenAI and shows them as a daily figure, a month-to-date total and a forecast for next month — next to the rest of your AI spend, in one pane.
Month to date · all providers
$1,385.00
- AWS
- OpenAI
- Anthropic
- Vercel
- Projection
Next month forecast
$2,780 – $3,190
Trend + fixed costs · high confidence
Monthly budget $3,500
On trackForecast uses 91% of budget
One admin key, used only to read costs
Tallypane reads your costs from OpenAI’s organization costs API. It needs an admin key, because that’s the only kind the costs API accepts, and it only ever uses it to read billing data.
Your app and its requests stay exactly as they are. Add the key once and Tallypane keeps your costs in sync, re-checking the last three days so late adjustments still land in the right place.
Spend by day and by model, month to date
See what OpenAI cost you each day and where it went, broken down by account and by model. A new feature, a model upgrade or a runaway batch job shows up the day it happens, not when the invoice arrives.
The month-to-date total sits next to this month’s projection, so you always know where the bill is heading.
Next month’s forecast, beside your other bills
Tallypane forecasts next month from your trend, with a low–high range and a clear flag when there isn’t enough history to trust it yet.
OpenAI is one row in a single table. Anthropic syncs today too, and OpenRouter, AWS, Google Cloud, GitHub and Vercel connectors come next. Switch between USD and BRL anywhere; conversions use the Central Bank of Brazil’s PTAX rate for each day.
Providers
Every bill you already pay, in one table.
OpenAI and Anthropic sync today. The rest are next — and anything without an API will come in as a manual or fixed cost.
- OpenAI· Production — Organization costs APISynced 2h ago$368.00
- Anthropic· Team — Admin API cost reportSynced 2h ago$352.00
- OpenRouter — Activity APISoon (coming soon)—
- AWS — Cost ExplorerSoon (coming soon)—
- Google Cloud — BigQuery billing exportSoon (coming soon)—
- GitHub — Billing usage APISoon (coming soon)—
- Vercel — Billing charges APISoon (coming soon)—
- Meta Muse — Manual entriesSoon (coming soon)—
- Supabase — Plan presetSoon (coming soon)—
Sample data. Hover a provider for where its numbers come from.
FAQ
Questions, answered.
Which OpenAI key does Tallypane need?
An OpenAI admin key, because the organization costs API only accepts admin keys. Tallypane uses it only to read cost data. Once saved, the key is encrypted with AES-256-GCM and never sent back to the browser — you only ever see its last four characters.
Does checking my OpenAI costs cost anything?
No. OpenAI’s costs API is free to query, so refreshing your numbers costs nothing and uses no tokens.
Can I see which models drive the bill?
Yes. Spend is broken down by account, model and service, so you can see which models and which accounts account for most of it.
How is this different from OpenAI’s own usage page?
OpenAI’s page shows OpenAI. Tallypane puts OpenAI next to Anthropic today, and next to your cloud and dev-tool bills as connectors land — one total and one forecast, in USD or BRL.
When does the hosted version open?
We’ll email everyone on the waitlist once — when it opens. That’s the only email you’ll get from the list.
Know the number before the invoice does.
Join the waitlist for hosted Tallypane. One email when it opens.