Measured Cloud documentation

Measured Cloud is a self-hostable data-quality platform. Point it at your warehouse, say what "good data" means for each table, and it checks on a schedule — telling you, in plain English, what broke, where, and why. The validation engine is the open-source Great Expectations; everything around it is Measured Cloud.

How it works

Every part of the product maps to one step of the same loop:

Step	What you do	Read
Data sources	Connect a database, warehouse, lakehouse table, or object store — directly, or through the Privacy Agent inside your network.	Data sources
Data assets	Add the tables you care about and profile them: row counts, null rates, distributions.	Data assets
Rules	Say what good data looks like, one check at a time — from a catalog, in plain English with AI, or as custom SQL. Rules live on their table.	Rules
Validations	Bundle the rules of related tables into one runnable unit with a health score.	Validations
Schedules	Run each validation on the cadence your data lands, in your organization's timezone.	Schedules
Execution runs	Watch runs fill in asset by asset, and stop one if you need to.	Execution runs
Results	See every rule's pass or fail, the failing rows, root-cause analysis, and an AI explanation.	Results
Alerts	Route failures by severity to Slack, email, Teams, PagerDuty, Jira, and more.	Alerts

Where to start

• Try it in five minutes. The Quickstart goes from git clone to a green run on your laptop.
• Walk the whole loop once. Run your first validation connects a source, adds rules, runs a validation, and schedules it.
• Add checks to your tables. Add rules to a data asset, then create and run a validation and schedule it.
• Hear about failures. Set up Slack or email alerts.
• Run it yourself. Self-host, then deploy to production with the configuration reference.
• Automate it. The REST API (with a full, generated OpenAPI reference), the Python SDK, the MCP server for your own LLM, and integrations for Airflow, Dagster, and dbt.

Help inside the app

Open Help from your user menu (bottom-left) for a searchable panel of these docs, or click the ? in a page's header to jump straight to the guide for that screen. See Get help in-app.

How these docs are organised

Section	What you find there
Getting started	The Quickstart, the first-validation walkthrough, self-hosting, and inviting your team.
Concepts	One page per object in the workflow above — what it is, how it behaves, what you can do with it.
How-to guides	Task recipes: goal, prerequisites, numbered steps, and how to verify.
Reference	The API, the dqc CLI, every configuration variable, the architecture, and the glossary.

Conventions

• Examples assume Measured Cloud lives at https://dq.example.com. Replace it with your own hostname.
• Placeholders in code — <org-id>, <ws-id>, dqk_... — stand for values from your own workspace.
• A page that describes something not yet shipped opens with a Status: planned callout, so you're never misled about what works today.