Introduction

What Datapace is

Datapace provides database teams with one tool to understand, run, and govern their databases across engines, and AI agents to improve their database workflows.

It builds one graph of every database you own: the entities, relationships, and semantics of the data, and the infrastructure around each database (instance class, storage, cost, performance, usage, freshness). Datapace infers that graph from what it reads, your experts confirm it, and the result is living documentation that people and AI systems can trust.

On that graph, Datapace's agents draft the routine work: documentation, cost optimization, performance, and migration mapping. Each proposal arrives with its evidence. Your team reviews and approves; nothing runs on its own, and every action is recorded.

Who it is for

• In-house database teams. DBAs, platform engineers, and data engineers at mid-size and large companies that run several engines, often with inherited or acquired estates, where what the data means lives in a few people's heads.
• Database managed-service providers. Teams that run Postgres, SQL Server, Oracle, and more for many clients, and need one way of working across every client's engines.
• Migration and ERP integrators. Teams handed decades of legacy data and expected to deliver a clean migration, who need the source mapped before a record moves.

How these docs are organised

• How it works walks the loop once, from connection to a recorded, approved change.
• The Concepts pages define each part of that loop: the graph, infer and confirm, the agents, policy, approval, and audit, governed context, and working across engines.
• The Guides describe how a team works with Datapace on a job: documenting an estate, reviewing a proposal, and mapping a migration.
• The Reference holds the glossary and the questions teams ask first.

Every page is also available as Markdown; see Docs for LLMs.

What is not documented yet

These pages describe how Datapace works. They are not an installation guide. How Datapace connects to a database, where it runs, and how data is handled are decided with each partner, and the reference for those surfaces (the collector, the API, the MCP endpoint) is published as each one ships.

If you want to see what Datapace reads on your own databases, book a call.