Who pays
Engineering managers at AI-heavy orgs, 5-50 devs
Weekly summary of your team AI spend + waste detection
Visibility first. Optimization second.
Tokens This Week gives engineering managers a weekly cost intelligence report across OpenAI and Anthropic: total spend, waste patterns, top expensive prompt flows, and cost-per-developer breakdown. You cannot optimize what you do not measure.
Niche
AI cost tools built for software teams
Cadence
Automated weekly email every Monday morning
Problem
Teams burn budget in the background with no weekly visibility. By the time finance flags AI spend, optimization is already late.
Solution
Connect provider keys once, then get a manager-ready weekly brief: spend totals, anomalies, waste insights, and highest-impact prompt fixes.
Why now
caveman + prompt-shrinker can optimize prompts, but teams need baseline visibility first. Tokens This Week closes that gap.
What you get every week
Total spend across OpenAI + Anthropic with trendline
Waste patterns detected with estimated dollars to recover
Top 5 expensive prompt flows to optimize first
Cost-per-developer breakdown for accountability
Reports are email-native, so managers can forward a concise weekly snapshot directly into standups or sprint planning.
Pricing
One plan. Clear accountability.
Tokens This Week is built for engineering managers running AI-heavy teams who need a weekly spend signal before rolling out optimization work.
$19
per team, per month
- Weekly spend digest across OpenAI + Anthropic
- Waste pattern detection with fix recommendations
- Top 5 expensive prompt flows to optimize first
- Cost-per-developer visibility for engineering managers
- Automated email report every Monday morning
Checkout is hosted by Stripe. Use the same work email for payment and dashboard unlock.
Start for $19/month1) Register team details
Save your team profile so we know who to activate after Stripe checkout.
2) Unlock paid dashboard
After checkout, enter the same billing email to activate your cookie-based access.
FAQ
Which APIs are supported?
Today we support OpenAI and Anthropic via read-only API keys. You can connect one or both providers, and the weekly report combines them into a single spend view.
How does waste detection work?
We analyze usage events for expensive patterns like oversized context windows, unbounded output verbosity, and premium-model overuse on lightweight jobs. Each finding includes a concrete fix recommendation.
Do you store model prompts?
We optimize for minimal retention: encrypted provider keys and derived spend analytics. If your provider usage feed does not expose prompt text, we report anonymized prompt flow fingerprints instead.
How quickly do we get value?
Most teams get an actionable first report in under one week: top expensive prompt flows, estimated waste, and a prioritized optimization queue.