Your keys.
Your code. Your call.

Security and trust here come down to control. The whole point of ShipItFam is that you stay in control — so the same goes for your data. You hold the trigger, and nothing important happens without you.

How we think about it

Control isn't a
feature. It's the point.

Every decision below comes from the same idea: the agent proposes, you dispose.

You bring your own key

The intelligence runs on your own Claude key, under your own account. We never sit between you and the model with a meter running.

Nothing ships without you

Every change waits for your approval. The worst case is a tap to reject — never a surprise in production.

Your code stays yours

Work happens in your own space, on your terms. Export it, leave any time, take everything with you. No hostages.

Model usage is on your account

Prompts go to Anthropic through your Claude key under their terms. ShipItFam does not use your code to train our own models. See Anthropic’s privacy policy for how they handle API data.

Session events in your workspace

Approvals, checkpoints, and runner events are appended to your project’s local event log so you can review what happened.

Control presets you choose

Pick how much autonomy the agent gets — from approving every step to reviewing whole features. You can tighten or loosen oversight per project.

Our promise

The short version,
in plain words

  • We don't bill you for tokens. Your key, your account, your usage — we can't even see the meter.
  • We don't ship behind your back. If you didn't approve it, it didn't happen.
  • We do not train on your code. Your work stays in your workspace; model calls use your Anthropic account.
  • We don't trap you. Cancel and walk out with everything you made.
Key on your account
You
Change proposed
Waiting
You tap approve
Shipped

Nothing in this list moves without you. That's the whole design.

Questions about trust?

Ask us
anything.

If something here isn't clear, we'd rather you ask than guess. We'll give you a straight answer.