Security model

Credentials, policy evaluation, network posture, and what a reviewer will ask you for.

10 min Owner: Security Updated 13 Sep 2026 v1.4

Least privilege by default

Read-only credentials, scoped to named schemas. Write access is granted per table and per action. There is no mode in which Ward holds an admin role on your warehouse.

In the console: Policies

Policy at query time

Row and column policy is evaluated when the query runs, for every caller. A rule enforced in the client is not enforced at all. This applies identically to a person in the console and to an agent.

In the console: Policies

Where your data sits

Your region, your retention window, and customer-managed keys on Enterprise. Object storage is read in place, so for lake sources the data does not move at all.

In the console: Architecture

What to send your reviewer

The architecture note, the DPA, the SOC 2 report and a certificate of insurance are available before you sign, not after. Ask and they arrive the same day.

In the console

The screens this guide walks through, captured from a running tenant.

app.getward.ai

A Cedar policy open in Ward’s editor, permitting the viewer role to query the sales.store_sales_daily model when the region is WE, with unit_cost, margin and extended fields marked hidden. — **Policy editor.** Row scope and hidden fields in five lines, with autocomplete over your own models and errors underlined as you type.

Ward’s Policies list for the Acme tenant: eight Cedar policies, seven permitting and one forbidding, each with an inline edit control. — **Policies.** Access is eight Cedar policies you can read in a sitting, not a permissions matrix nobody opens.

Ward’s access log: every authorization decision recorded for the tenant, timestamped to the second in UTC, with outcome, action, resource, actor, and the milliseconds each decision took. — **Access log.** Every authorization decision, allowed or denied, with the actor and the latency. This is the page the security review asks for.

The Sources list in Ward, showing five connected systems with their type: two Parquet exports, an HTTP API, and two API connectors, above a New source button. — **Sources.** Every external system the lake reads, in one list, with the connector each one came in through.

Ward’s Connections page listing the database backends that sources, models and pipelines reference: a managed lake and a configured PostgreSQL connection. — **Connections.** The backends everything else points at. Credentials configured once, referenced by name after that.

Ward’s schema diagram at full screen: 29 model tables with every field typed and the foreign keys drawn between them, zoomed out to 25 per cent to fit. — **Schema diagram.** The entity map, generated from the models themselves and exportable as source.

PreviousSemantic layer