See your team's Claude spend — per engineer — and lock $199/mo forever
The first 10 engineering teams to sign get white-glove TokenShield install, a shared dashboard that attributes every Claude dollar to an engineer, model, and project, weekly reports for your CFO, and every optimization processor as it ships — at a rate that never goes up.
$199/mo flat · up to 10 engineers · cancel anytime · 30-day money-back · no setup fee
First, you can finally see it
A typical 20-engineer team using Claude Code heavily.
You can't cut what you can't see. We give you the attribution first — that alone is worth the seat to a team with a finance lead asking questions. Optimization is the opt-in second act, measured against the baseline we capture in Week 1. We don't quote a savings percentage we can't back up with your own data.
What you get
Everything here is included in the $199/mo flat retainer — no upsells, no setup fee.
White-glove install (30-min Zoom)
We get TokenShield running on every engineer's machine, route Claude Code through the proxy, and verify the ledger is recording — same day.
Weekly spend report
Every Friday: a PDF showing your team's measured Claude spend per engineer, per model, per project. Forward straight to your CFO — it answers 'wait, where IS the money going?'
Every processor as it ships
conversation-dedup, result-cache, diff-file-reads, context-summarizer — each one becomes available to your team the day it goes GA, with its effect measured in your own dashboard. No upgrade pricing.
Dedicated Slack/Discord channel
Shared channel with our engineering team. Bug in your config? Ask. Want a custom processor? We listen. Direct line, not a ticket queue.
Founding rate locked forever
$199/mo flat for up to 10 engineers, forever, even after we move to per-seat retail pricing.
30-day money-back
Cancel in the first 30 days for any reason and we refund the month — no questions, no savings target to argue about. Month-to-month after that.
Timeline
Apply + kickoff scheduled
- Submit the form below — we reply within 24 hours
- 30-minute kickoff Zoom scheduled within 3 business days
- We send the Founding Team agreement (1-page PDF) and your first month's invoice
White-glove install + baseline
- Live install session: every engineer's machine gets TokenShield
- Ledger verified recording in your private CuratedMCP dashboard
- Baseline established — we capture exactly what your team spends today, attributed
- First weekly spend report delivered Friday
Full visibility, measured optimization
- Per-engineer, per-model, per-project spend visible in your shared dashboard
- Each optimization processor activates for your team — with its real effect measured against baseline
- Weekly PDF spend reports
- Bi-weekly 15-min check-ins (or async — your call), custom requests prioritized
Continue at the founding rate
- Stay at $199/mo flat (locked forever) or cancel anytime — month-to-month
- You keep the data + reports from the pilot either way
Good fit if you're
- A 5–100 engineer team using Claude Code, Cursor, or Windsurf daily
- Spending $1,000+/mo on Anthropic API across your team
- Want per-engineer + per-project visibility into AI spend
- Have a finance team that's started asking questions
Not a fit if
- ×You only use Claude.ai chat (not coding tools)
- ×Your team is < 5 people and spend is under $500/mo (free CLI is plenty)
- ×You can't change your team's
ANTHROPIC_BASE_URLenv var - ×You need on-prem only, no telemetry — we route through CuratedMCP cloud for the dashboard
Quick FAQ
Why $199/mo if I can install the CLI myself for free?
The free CLI is perfect for one engineer. For a team, the retainer buys you: white-glove rollout to every machine in one Zoom, a shared dashboard that attributes spend per engineer/project, weekly CFO-ready reports, and a baseline that becomes your measurement yardstick. The visibility is the product; the optimization is the opt-in upside.
How much will we save?
We won't quote a number we can't back up with your own data. On prompt-cached workloads (Claude Code's default) some repeated context is already discounted, so the optimization win is workload-dependent. We capture your baseline in Week 1 and report measured effect against it — you decide with real numbers, not a marketing percentage.
What if my team uses Pro/Max subscriptions instead of API keys?
TokenShield currently routes Anthropic API key traffic. If your team is on Pro/Max OAuth, the pilot includes helping you migrate the heaviest users to API access (which usually saves money on its own at scale).
What happens to my data?
Prompt content never leaves your machines. Aggregate metrics (request counts, token counts, model, duration, cost) sync to your private dashboard at curatedmcp.com. Deletable on request, exportable as CSV anytime.
Is this a contract?
No. $199/mo is month-to-month, cancel anytime, with a 30-day money-back on your first month. The only commitment is ours: we hold the founding rate for you forever.
Can my CISO/legal review your DPA?
Yes. Email [email protected] and we send the DPA template, sub-processor list, and SOC 2 roadmap summary same-day.
Apply for the Founding Team
Spots fill in the order applications arrive. We review every submission personally and reply within 24 hours.
Questions before applying? Email [email protected] — we read every message.