Arcane
      db      \`7MM"""Mq.   .g8"""bgd     db      \`7MN.   \`7MF'\`7MM"""YMM
     ;MM:       MM   \`MM..dP'     \`M    ;MM:       MMN.    M    MM    \`7
    ,V^MM.      MM   ,M9 dM'       \`   ,V^MM.      M YMb   M    MM   d
   ,M  \`MM      MMmmdM9  MM           ,M  \`MM      M  \`MN. M    MMmmMM
   AbmmmqMA     MM  YM.  MM.          AbmmmqMA     M   \`MM.M    MM   Y  ,
  A'     VML    MM   \`Mb.\`Mb.     ,' A'     VML    M     YMM    MM     ,M
.AMA.   .AMMA..JMML. .JMM. \`"bmmmd'.AMA.   .AMMA..JML.    YM  .JMMmmmmMMM

Govern as fast as employees build

Talk to founders
see how it works
1.0 · The shift

The risk isn't just Shadow AI. It's the speed.

Employees build and ship AI workflows faster than any review process can keep up. Arcane works as fast as the threat model.

PLATFORM APPROVED
The tool is allowed.
But the app built inside it may be invisible.
WORKFLOW CREATED
Data, secrets, and tools get connected.
Often outside formal SDLC.
SECURITY REVIEW
Manual review cannot keep up.
Arcane handles the drudge and escalates what matters.
2.0 · The platform

Governance that outpaces development.

Enforce now, watch always, tighten over time. Arcane monitors behavior 24/7, drafts policy changes, and queues meaningful exceptions for approval so security stays in control without turning every workflow into a ticket.

arcane / policy loop
01 · Observe
30d traffic + denials + drift
active
02 · Propose
Posture agent drafts change
03 · Review
Operator inspects + replays
04 · Promote
Active policy set updated
watching traffic
tool calls · 30d
18,432
denials · 30d
47
agents drifted
3
baseline → proposal → operator → activecontinuous
arcane / proposal · 1 of 3
proposed by posture agent

allow research-agent to read drive.files when delegated by an engineering-org user

reasoning

research-agent attempted drive.files.read 47 times in the last 30 days — all from engineering users. No matching rule. Promoting shadow → active would convert 47 reviews to allows.

impact if applied
+47 newly allowed362 unchanged
cedar
permit (
  principal in Group::"research-agent",
  action == Action::"read",
  resource in ResourceType::"drive.files"
) when {
  principal.delegated_by.org == "engineering"
};
posture agent · proposal queuepolicy stays in shadow until you promote
3.0 · How it works

Three calls. Identity to enforcement.

Bind once at boot. Mint per task. Enforce on every call. The whole control plane fits behind three SDK methods.

01

Bind identity

Arcane composes the user from your IdP, the agent from your registry, and the workload from runtime attestation into a single signed principal.

arcane.tsSDK
// At agent boot
const principal = await arcane.bind({
  user: req.identity,           // verified IdP token
  agent: { id: "release-bot", version: "1.2" },
  workload: await runtime.attest(),
});
// → { principalId, boundAt, expiresAt }
02

Issue a task-bound grant

Each grant names the task, the resources it covers, and the actions allowed. Short TTLs replace standing access; nothing else carries.

arcane.tsSDK
// Per-task grant
const grant = await arcane.issue({
  principal,
  task: "rotate-staging-keys",
  resources: ["aws:iam:role/staging-*", "vault:kv/staging/*"],
  actions: ["iam:UpdateAccessKey", "kv:read"],
  ttl: "5m",
});
// → { token, exp, jti }
03

Enforce on every call

SDK at the call site, or transparent proxy in front of any service. Same policy, same signed evidence — evaluation happens inline, not out of band.

arcane.tsSDK
// Guarded call — policy evaluates before the SDK forwards
await arcane.guard(grant, () =>
  aws.iam.updateAccessKey({
    UserName: "ci-deploy",
    Status: "Inactive",
  }),
);
// decision: allow · signed · evidenceId=ev_3f9c…
4.0 · The console

Every decision, in plain view.

The same record that flows to your audit log shows up live — who, what, why, and which rule matched. Filter by agent, task, or outcome.

arcane / decisions
time
outcome
agent · task
resource
showing 6 of 1,284live
5.0 · Get started

One conversation.
Enable your security team.

Talk to founders