Skip to content

NEWUnify Agents (MCP): governed AI access to finance data. Early access.See Unify Agents

GLOSSARY

What is change data capture (CDC)?

DEFINITION

Change data capture detects inserts, updates and deletes in a source system and emits them as a stream of change events.

Also called: CDC, Change capture

How it works

Database CDC reads the transaction log, such as the Postgres write-ahead log, and publishes each row change in order. For SaaS platforms without log access, CDC is built from change feeds, webhooks, modified-since queries or periodic comparisons of snapshots. Deletes are the hard part: many APIs simply stop returning a deleted record, which only a comparison can detect.

Example

A ledger API has no deleted-records endpoint. A nightly comparison finds that invoice INV-2207 disappeared, and emits a delete event so downstream systems remove it too.

In integrations

Unify Sync is designed to combine native webhooks, modified-since polling and snapshot comparison per platform, so changes on the Unify Graph, deletes included, arrive as one consistent stream of events and webhooks. Reliable change capture is what keeps a stored copy trustworthy. Without delete detection and ordering guarantees, small errors accumulate until the copy no longer matches the source.

Related terms

Related products and solutions

Get started

Connect once. Build everything else on top.

Request sandbox access, or book 30 minutes with an engineer to map your integrations.

Free sandbox access. No credit card. Reply within one business day.