We use analytics cookies to understand site traffic. You can accept or decline.
00 / THE PLATFORM

The platform behind governed AI operations.

Droco is not a single feature. It is the control plane, the console, and the record layer for running AI and infrastructure work in production — with you owning the result end to end.

01 / SURFACES

One platform, seven parts you own.

Each part is inspectable, exportable, and runs where your policy allows. Together they cover the whole path from request to proof.

01

Control plane

One core that knows who can do what — identity, permissions, policy, and audit history, all in one inspectable place.

02

Operator console

A real dashboard to see what's running, approve sensitive work, and pull the records — not just a status page.

03

Policy & approvals

Risky actions wait for a named human. Every approval is tied to a person, logged, and expires on its own.

04

Tool registry

Everything automation is allowed to do is written down, reviewed, and bounded before it touches a real system.

05

Observability

Metrics, logs, and error tracking built in — so you see the health of the system instead of guessing at it.

06

Evidence & exports

Every run produces a portable record — request, plan, approval, execution, proof — that stays yours.

07

Self-hostable & federated

Runs where your rules allow — your servers or your cloud. Open formats, so leaving is always an option.

02 / THE GOVERNED LOOP

How the platform governs every action.

Anything sensitive runs the same loop: a clear request, a specific approval, and a record you keep. This is the spine that makes the rest of the platform trustworthy.

INTENTStart with a clear request.

Before sensitive tools run, Droco captures the requested change, its expected impact, and what is out of scope. The request is tied to a person, agent, or service identity.

{
  "summary": "rotate prod database credentials",
  "blast_radius": "core-prod",
  "tools_required": ["vault.kv.put", "nomad.job.restart"],
  "approver_required": true,
  "expected_outcome": "services reconnect with rotated credentials"
}

APPROVALSpecific, time-bound, and reversible.

Approvals are tied to identity, policy, and environment. Each gate records who approved it, when it expires, and what it allowed.

approval.id          : a47f-2826
approval.intent      : intent#7821
approval.approver    : platform-owner · authorized approver
approval.policy      : prod.rotate.requires_cto
approval.granted_at  : 2026-04-25T14:03:11Z
approval.expires_at  : 2026-04-25T14:33:11Z
approval.revocable   : true

EVIDENCEKeep proof you can inspect later.

Each governed run produces a bundle with the request, plan, approval, execution log, and related artifacts. Your team can export it, review it, and keep it with the rest of your records.

loading proof…

Want to see it on a real workflow?

Request a working session →