Change data capture for application teams
Driftbase reads the Postgres write-ahead log and applies every committed insert, update and delete to your search index and cache, in commit order, usually in under a second. You delete the cron job, the dual writes and the “reindex everything” button.
brew install driftbasedev/tap/drift
Sinks that ship today
Driftbase attaches to a logical replication slot with the built-in pgoutput plugin. No triggers, no extension to install, no extra load on your hot tables. It works on RDS, Aurora, Cloud SQL, Neon, Supabase and plain Postgres 12 and up.
Describe each index as a query: joins, computed columns, nested arrays. When anorder_items row changes, Driftbase works out which orders documents it belongs to and rewrites only those.
New index? Changed mapping? drift backfill copies existing rows in keyset chunks while live changes keep flowing, then hands over without a gap or a duplicate.
Every pipeline reports end-to-end lag, slot size and per-sink error rates, and exports them to Prometheus and Datadog. Alerts fire before the slot can fill your disk.
“We had a nightly reindex and three different places that wrote to Elasticsearch. Driftbase replaced all of it in an afternoon, and our support queue about stale search results went to zero.”
Run the CLI against your own infrastructure or let us host the pipeline. Either way, the Free plan covers 1,000,000 change events a month with no card.
Create a free workspace