Every Data Transport Pattern. One Platform.

ETL & ELT

Classic extract-transform-load, or load-first with in-warehouse transformation.

Representative connectors See All 400+ Connectors

Key features

• Runs on a schedule, from minutes to daily, for classic batch extract-transform-load or load-first ELT.
• Extracts structured tables and API objects, and can add dataddo_hash and extraction-timestamp metadata columns for clean upserts and history.
• Supports append, replace, and upsert write modes, so you control exactly how records land in the destination.
• Lets you select only the columns you need or hash sensitive ones, keeping PII out of the destination.
• Can run a full data re-sync at any time by widening the date window to backfill historical records.

Change Data Capture (CDC)

Real-time, low-latency replication that tracks row-level changes as they happen.

Representative connectors See All 400+ Connectors

Key features

• Replicates changes in near real-time with log-based CDC, reading committed changes straight from the database transaction log.
• Captures inserts, updates, and deletes - the only pattern that reflects deletions - while putting very low load on the source.
• Writes with append, replace, or upsert modes and can add dataddo_hash and extraction-timestamp metadata columns.
• Supports a full initial re-sync that can be split across several scheduled runs for very large tables.

Data Streaming

Continuous, event-driven pipelines for time-sensitive and AI-ready data workloads.

Representative connectors See All 400+ Connectors

Key features

• Ingests data continuously from event sources such as Kafka topics and inbound webhooks, rather than on a fixed schedule.
• Runs through the same governed flows as batch data, so schema, access, and lineage are managed identically.
• Lands the stream directly in warehouses, AI context stores, or downstream apps for time-sensitive workloads.
• Supports append and upsert write modes and can re-sync the full dataset whenever you need to.

Reverse ETL

Activate your data: push curated, governed records from your AI agents or warehouse back into CRMs, operational systems, and the frontier apps where your teams act on it.

Representative connectors See All 400+ Connectors

Key features

• Pushes curated data from your warehouse or source back into business apps such as CRMs, ad platforms, and operational systems.
• Maps source columns to destination fields and supports append, replace, and upsert write modes.
• Keys upserts on dataddo_hash, so existing records are updated in place instead of duplicated.
• Runs on a schedule to keep scores, segments, and AI-generated attributes current, and can re-sync the full dataset on demand.

Batch File Delivery

Structured delivery of datasets via files to S3, SFTP, or any storage target.

Representative connectors See All 400+ Connectors

Key features

• Delivers and reads structured files in CSV, JSON, XML, or Parquet, plain or GZIP-compressed.
• Follows date-stamped file names automatically with dynamic date placeholders, on a schedule from minutes to daily.
• Shapes or masks data before delivery with an editable, jq-capable transformation step, handy for dropping or transforming PII.
• Uses the replace write mode to mirror a file in place or append to build history, and supports a full re-sync.

Zero-Copy with Apache Arrow

High-performance, in-memory data sharing without serialization overhead, purpose-built for AI and analytics pipelines.

Representative connectors See All 400+ Connectors

Key features

• Shares data in-memory using Apache Arrow's columnar format, avoiding the serialization overhead of row-by-row transfer.
• Purpose-built for AI and analytics pipelines where large datasets move between processing steps at in-memory speed.
• Reads directly into Python, pandas, or notebooks for ML feature engineering and experimentation.