The whole replication stack,
in one binary.
TwinCommit is a log-based change data capture platform. It reads the database's own transaction log, assembles complete transactions and applies them exactly once, in commit order, to the platforms your teams use. No agents, no manager, no triggers.
1 binary
no JVM, no agents, no app server
5 min
from download to first replicated row
0 triggers
or capture tables on the source
Exactly once
commit-ordered delivery
Platform
Capture. Deliver. Operate.
Capture
- ✓Native log readers for Oracle, MySQL, SQL Server, PostgreSQL, MongoDB, Db2, Cassandra, MariaDB, SAP HANA
- ✓No triggers, CD tables or capture agents
- ✓Savepoint and rollback aware
- ✓Archive-log aware: survives log rotation
Deliver
- ✓Commit-ordered apply, one transaction at a time
- ✓Checkpoint committed with the data, so exactly-once
- ✓Native MERGE, COPY, Snowpipe Streaming and Kafka transactions
- ✓Spill-to-disk for very large transactions
Operate
- ✓Plain YAML pipelines with ${ENV} secrets
- ✓Prometheus /metrics and JSON /status
- ✓Table-level fault isolation and one-table re-snapshot
- ✓MCP server for Claude, Copilot and Cursor
Sources
Oracle Database
Redo and archive log mining with no LogMiner session overhead.
MySQL
Row-based binlog capture with GTID positioning.
Microsoft SQL Server
Transaction-log capture without enabling SQL Server CDC jobs.
PostgreSQL
Logical decoding in, high-speed COPY + upsert out.
MongoDB
Change streams in, transactional bulk upserts out.
IBM Db2
Db2 LUW and Db2 for z/OS, read straight from the recovery log.
Apache Cassandra
Commit-log CDC from every node, deduplicated into one ordered stream.
MariaDB
Binlog capture with MariaDB GTIDs, Galera-aware.
SAP HANA
Redo-log capture for SAP HANA, without trigger-based replication.
Targets
Oracle Database
Redo and archive log mining with no LogMiner session overhead.
MySQL
Row-based binlog capture with GTID positioning.
Microsoft SQL Server
Transaction-log capture without enabling SQL Server CDC jobs.
PostgreSQL
Logical decoding in, high-speed COPY + upsert out.
MongoDB
Change streams in, transactional bulk upserts out.
Snowflake
Snowpipe Streaming ingest with a MERGE that preserves deletes.
Redis
Keep caches perfectly in sync with the system of record.
IBM Db2
Db2 LUW and Db2 for z/OS, read straight from the recovery log.
Elasticsearch
Search indexes that are never stale and never out of order.
Databricks
Stream committed transactions into Delta Lake tables in Unity Catalog.
Apache Cassandra
Commit-log CDC from every node, deduplicated into one ordered stream.
MariaDB
Binlog capture with MariaDB GTIDs, Galera-aware.
SAP HANA
Redo-log capture for SAP HANA, without trigger-based replication.
Amazon Redshift
S3-staged COPY plus MERGE for a warehouse that mirrors production.
Google BigQuery
Storage Write API ingest with MERGE-based upserts and deletes.
Azure Synapse Analytics
ADLS-staged COPY INTO plus MERGE for dedicated SQL pools.
Azure Data Lake Storage
Delta Lake or Iceberg tables in ADLS Gen2, ready for Synapse, Fabric and Databricks.
Amazon S3
Iceberg tables in S3, registered in Glue or S3 Tables, queryable from Athena.
Google Cloud Storage
Iceberg and Delta tables in GCS for BigLake, BigQuery and Dataproc.
Apache Spark
A Structured Streaming source for exactly-once change feeds.
Apache Hive
ACID MERGE into Hive transactional tables, or Iceberg tables via the metastore.
Apache Kafka
Transactional producers, one topic per table, keys preserved.
Replicate your first table in five minutes.
Download the free Developer edition, or have one of our replication engineers walk you through a migration plan.