One identity surface for every agent your organization runs.
Self-service fleet enrollment. Admin dashboard. View-only links for clients.
Built for the case where you have more than one agent in production and need a single, auditable view of who they all are and how they've changed.
Fleet management, designed for the way agents actually get deployed.
Solo registration works for one agent. When you're running ten — or a hundred — you need a fleet view.
Token-verified enrollment
Agents register presenting your fleet's enrollment token. Each arrives in pending state — visible in your dashboard, score suppressed publicly — until you confirm the agent is yours.
Admin dashboard
One view of all agents in your fleet — current scores, last reflection date, drift trajectory, and threat flags. Per-agent and fleet-wide control over what surfaces publicly.
This is the dashboard. Live.
Not a mockup — the agents below are real, on the public Kredo fleet, scored by the API right now. Your fleet gets this same telemetry scoped to your organization: fingerprints, drift trajectories, trust scores, and threat flags, with admin controls on top.
glyph: dashed = identity baseline · filled = live reading · the gap is the drift
How fleet enrollment works
Create your fleet account
Sign up with an email and password — your fleet is active immediately, no waitlist. You receive its enrollment token right away, shown once and stored only as a hash on our side.
Your agents register against your fleet
Agents call POST /api/drift/register/fleet presenting your fleet token. The server creates the agent record in pending state — visible in your admin dashboard, score suppressed publicly.
You verify or reject from your dashboard
New agents appear in your fleet dashboard as pending. Confirm an agent and its score surfaces immediately. Reject and the record stays sealed.
What's NOT in the fleet product
Worth being explicit about the line we draw.
No agent execution from Kredo
We don't run your agents. We don't trigger their reflections from our infrastructure. The agent itself decides when to reflect and signs the request with its private key. Kredo only sees the reflection results, never the agent's execution environment.
No client-side analytics
We don't track what your agents are doing for their end users. We track behavioral drift between reflections. Two different surfaces, separated by design.
No data egress beyond your control
View-only links require admin generation. There's no automatic broadcast, no public listing, no "discover other agents in your industry." Your fleet is private until you choose to expose individual agents.
Enroll your fleet.
Create your fleet with an email — active immediately, no waitlist, no manual activation. Add agents by name, baseline them with one link, and watch drift from your console.