Public docs

Commercial docs that explain the product before the pilot starts.

QueryRook keeps operator-only implementation evidence inside the app workspace. These public docs explain the buyer-facing workflow, AI boundary, MCP strategy, and security posture without dropping visitors into the customer dashboard.

User guide

How a buyer uses QueryRook.

The public guide is intentionally buyer-facing: it explains the operating model without exposing private tenant data, admin-only controls, or internal deployment tasks.

01

Enroll a Postgres target with least-privilege credentials and explicit environment labels.

02

Let QueryRook build workload evidence before any production-facing recommendation is promoted.

03

Review proof packets, risk summaries, and rollback posture before approving a change.

04

Export the buyer packet for diligence, pilot review, or internal change-control records.

AI backbone

Bounded AI, not mystery automation.

QueryRook can use AI to reason over evidence, but the product boundary is built around scoped context.

Context boundary
The AI receives scoped evidence summaries, workload metadata, and approved task context.
Secret boundary
Provider keys, DSNs, and customer secrets stay out of prompts and are handled by platform services.
Action boundary
Autonomous work is constrained by policies, approvals, dry-runs, proof checks, and rollback readiness.
MCP strategy

Evidence tools with tenant scope.

MCP is useful only when the tool boundary is narrow enough for serious customers to audit. QueryRook's strategy is to expose inspectable evidence and governed actions, not open-ended database control.

01

Tenant tools

MCP-style tools are scoped to the tenant, role, and active workflow.

02

Evidence first

Tools expose proof packets, run status, catalog metadata, and safe summaries before mutation paths.

03

No raw reacharound

The product path is designed to avoid handing a model direct, unconstrained database access.

Security doc

Buyer diligence lives in the security packet.

The security packet covers control boundaries, evidence retention, deployment posture, AI/secret separation, and the materials a pilot buyer can request before production access.