Every Snowflake platform needs the same three things: proof the numbers are right, proof the governance holds, and an admin layer that doesn’t drift. We built them once, properly, as products that run entirely inside your Snowflake account — as your roles, with nothing leaving.
Data quality and reconciliation. Bronze, Silver and Gold reconcile every day, natively, with the row-level difference when they don’t.
About Vigil →Governance evidence. Classification, tag-driven policies, drift detection and audit evidence packs — configured on day one, not after a six-month project.
About Pramana →Snowflake administration. RBAC, Iceberg tables and disaster recovery, provisioned declaratively and reviewable in a pull request.
About Dhruva →What it is. Layer-to-layer reconciliation and data quality checks, built entirely on Snowflake: Data Metric Functions, serverless tasks and alerts, and a Streamlit console. Bronze, Silver and Gold reconcile every day without a dbt run or an external tool.
Why it matters. When the Gold number doesn’t match the source, you want to know before the business does — with the row-level difference, not a red light.
How it fits. Vigil emits its results as events to Snowflake’s account event table, so they show up in Snowflake Observe alongside everything else. Reconciliation results feed Pramana as audit evidence.
What it is. Classification, tag-driven masking and row-access policies, drift detection and audit evidence packs. A policy pack you load and configure on day one, not a governance project you run for six months.
Why it matters. When the auditor says “show me”, you hand over an evidence pack, not a spreadsheet and a promise.
How it fits. Pramana consumes Vigil’s reconciliation results as evidence and works alongside the roles and policies Dhruva provisions.
What it is. RBAC, Iceberg tables and disaster recovery, provisioned declaratively — Snowflake’s own declarative change management first, Terraform where that doesn’t reach yet.
Why it matters. Platform administration that lives in version control and is reviewable in a pull request, not scattered across worksheets and tribal knowledge.
How it fits. Dhruva provisions what Pramana governs and Vigil checks. The three are designed to connect.
Dhruva provisions the platform. Vigil proves the numbers reconcile. Pramana proves the governance holds — and keeps the evidence. Use one, or all three.
Dhruva stands up roles, storage and recovery from declared intent, so the platform starts governed instead of becoming governed later.
Vigil reconciles every layer to the one before it and surfaces differences as rows, so a wrong number is caught before it reaches a report.
Pramana records what is classified, what is protected and what has drifted — and hands you the evidence pack when someone asks.
Vigil and Pramana are being packaged as Snowflake Native Apps for the Snowflake Marketplace. All three products are available now for early-access and private deployments into your own Snowflake account, with our team alongside you.
We’ll deploy into a sandbox of yours and walk through it with your team — your data, your roles, nothing leaving.
Request a Walkthrough