Architecture
You chose Bedrock so your data would stay yours.
Enterprises do not run Claude through Bedrock for the pricing. They run it through Bedrock because the model runs inside their own AWS environment, prompts and completions are not used to train the underlying models, and nothing is handed to a third-party SaaS boundary.
A SaaS spend tool in front of it undoes the decision you spent six months getting approved. So TokenLock is not one. The stack deploys into your account, the console runs browser-to-your-own-API, and we operate nothing in the path.
How it works
Your data does not leave your AWS account.
You chose Bedrock so your data would stay yours. A SaaS spend tool in front of it undoes the decision you spent six months getting approved. TokenLock deploys into your account, and the console runs browser-to-your-own-API. We operate nothing in the path.
Your people
Still your account
There is no vendor to assess
Because there is no vendor in the data path. Security review is of your own infrastructure, which you already have a process for.
Residency is recorded, not assumed
Cross-region inference profiles mean a request is not automatically served from the region it was sent to — a global. profile may be served from any geography. TokenLock prefers regional profiles and records which one actually served each request.
What we can see: one number
Nothing about a request leaves your account: not prompts, not responses, not who spent what. There is no telemetry, no shared control plane, and no account of ours in your VPC.
The one thing we read is the number we invoice against. Our billing run asks your stack, over a shared key, for a count of active users, per-day counts and a request total — nothing identifying, and a test fails if that ever changes.
| Component | Runs in | Holds |
|---|---|---|
| tokenlock-proxy | Your VPC | Nothing at rest |
| tokenlock-console | Your VPC | Served to the admin browser |
| tokenlock-db | Your account | Usage, budgets, conversations if retained |
| bedrock | Your account | Model inference |
| identity provider | Your tenant | Who your people are |
| billing endpoint | Your stack | Answers our billing run with counts only |
Residency, including the part nobody raises
Cross-region inference profiles mean a request is not automatically served from the region it was sent to. A global. profile may be served from any geography.
TokenLock resolves regional profiles in preference to global ones, and records which profile actually served each request. Residency becomes a recorded fact rather than an assumption.
Almost nobody asks about this. It is on the page because it is the part your security team will check.
Corporate sign-in
Entra ID and Google Workspace federation. Department comes from the directory, so team reporting needs no extra configuration.
Administrator-only console
Every console route requires the administrator group. Staff do not browse each other's spend.
Retention you choose
Off, 7d, 30d, 90d, 1y or indefinite. What that means for staff →