Blog

Notes from the boundary

Writing on request-level authorization, machine identity, and what it takes to let AI agents touch production APIs safely.

01 Latest

13 min

iGaming analytics: prove what touched the data, not just who logged in

Gaming operators love what a shared analytics platform gives them — and fear what they can't verify. Six places where request-bound proofs turn 'trust us' into 'here is the evidence.'

Read
9 min

Authorization without a round trip: where the latency actually goes

The first question every architect asks about a boundary in front of their API is what it costs them in milliseconds. The honest answer depends entirely on which parts of the system participate in a request — and for us, the answer is one.

Read
6 min

Why API keys keep leaking — and why rotation doesn't fix it

Rotation shortens the blast radius of a leaked key, but the key was already a standing credential. The structural fix is to stop sending one.

Read
5 min

AI agents act in milliseconds. Your authorization should too.

Agents chain tool calls faster than any human review loop. If a request can't prove itself at the boundary, your agent framework is running on trust.

Read
7 min

Inbound vs. outbound: two halves of machine identity

Workload identity platforms govern what your services can reach. A boundary governs what can reach your services. They're complementary — and you probably need both.

Read
5 min

Valid once: what single-use proofs buy you

Replay resistance, theft-worthlessness, and a clean audit trail — all from one property: a proof that works exactly one time.

Read