Three parts. One calm workflow.
An agent that does the careful work, a search screen for your team, and an optional dashboard to see the results. The first two run on your side.
Moves old rows, carefully.
A small program you run next to your database. It finds old data, copies it to your bucket, checks the copy, and removes rows only when you allow it.
- Finds the tables and collections with the most old data
- Previews exactly which rows a rule would move
- Works in small batches so your database stays responsive
- Compares each row with the archive before deleting it
- Resumes safely after a crash or restart
- Runs on a schedule, or once when you choose
A search box for the past.
When a customer asks about an old order, your support team should not need to file a ticket with engineering. The Gateway lets approved people find and export archived records in their browser.
- Filter by any column and open the full record
- Export to CSV, JSON Lines or Parquet
- Company sign-in with OpenID Connect
- Access granted per dataset and per column
- Searches, record views and downloads are logged
See the savings, not the data.
An optional dashboard for the whole team. Your agents send a short summary of each job, never the rows themselves, so you can track progress across databases in one place.
- Status of every job across every agent
- Table and index size, before and after
- Measured savings shown apart from estimates
- Read-only: it cannot start, stop or change jobs
Example figures for illustration.
Start small. Grow with confidence.
Most teams begin with one large, stable table and run copy and check only. Once the archive looks right in the Gateway, they turn on deletion for that table and add more over time.
Go deeper on your setup.
Let your database keep only what it needs.
We are working closely with a small group of PostgreSQL teams. Tell us about your largest tables and we will help you run the first archive.