Context Graph
Data Sync keeps the Context Graph current, so agents work from live data.
Coming soon — Data Migration Agent: agentic data migration, done in weeks, not months See how it works
Blunox Data Sync is real-time replication. It snapshots the tables you choose, then streams every insert, update and delete from the database log, with no triggers and no polling.
Set up a sync from the console. Data Sync runs the capture, the snapshot handover and the retries.
A source and a sink, each tested before you save. Credentials are Blunox-managed or your own.
Rename, drop, derive and fan in. Override a primary key where the source has none.
A parallel snapshot loads what is there, then the stream takes over from the log without a gap.
Status, lag and run history for every sync, with alerts when something needs you. When the source changes shape, your policy decides: widen the sink, wait for approval, or stop.
Reads committed changes from the database log: logical decoding on Postgres, CDC on SQL Server, change streams on Spanner and MongoDB.
Idempotent writes at the sink plus replay from the last durable checkpoint. The docs state the delivery guarantee for each sink.
Each change is classified. Widen the sink, hold it for approval, or fail the sync: you set the policy.
Run it on Blunox, or put the runner in your own cloud. Runners only make outbound connections.
Envelope-encrypted credentials, role-based access, tenant isolation and an audit log.
Start, stop and restart a sync, or re-snapshot a table, without rebuilding anything.
Connect a database and a warehouse, and watch the snapshot hand over to the stream.