Integration

Zembula + Amazon S3

Amazon S3 stores any file your team writes, such as a CSV product feed, and Zembula can read that file. Your team saves its access details in Zembula.

It's our business to grow yours

Book a demo

Zembula has built this integration.

What Zembula can read from Amazon S3

Zembula has a saved Amazon S3 connection. Your team sets it up by pasting its access details into Zembula, and Zembula can then read files from your bucket. S3 stores each file as an object, and Zembula can find an object by its bucket name and its object key. A common object is a CSV product feed.

An export file with person rows or product rows works the same way. S3 also holds tabular data in Apache Parquet format, kept in a table bucket. Your team decides the file layout, because the layout of each file belongs to you. Zembula can read an object through the S3 REST API, and it can take the file as a file feed.

Your email service provider can send the lookup value as a merge tag in the image URL. S3 serves at least 5,500 read requests per second for each partitioned prefix. Zembula can pick the newest file under a prefix as the current catalog.

DataDetailLooked up by
ObjectA file in a bucket, such as a CSV feed.bucket name and object key
TableTabular data in Apache Parquet format, stored in a table bucket.set at setup
Export file rowsPerson or product rows in an export file that your team writes.bucket name and object key

How the data arrives

  • API: Zembula can read an object from your bucket through the S3 REST API with the access details you saved.
  • File feed: Your team writes a feed file to S3, and Zembula can read it as the product or person source.
  • ESP merge tags: Your email service provider sends the lookup value as a merge tag in the image URL.

What you need

  • Access details for your Amazon S3 bucket, pasted into Zembula.
  • The bucket name and the object key of each file Zembula reads.
  • A lookup value for each person in your email service provider, sent as a merge tag.

Limits to plan for

LimitWhat it means
At least 3,500 write requests or 5,500 read requests per second for each partitioned prefix.Each prefix serves at least this many reads each second.

Ideas for email

  • Read a nightly product feed file to pick the items for an email.
  • Fill merge fields from a CSV export of person rows.
  • Pick the current catalog from the latest file under a prefix.

Last checked: 19 November 2026.

Performance marketing dials, inside email. For the first time.

See how we do it by booking a demo.