The high-performance transaction log processor

From commit to action.

Connect a source for its selected mode. Remac operates the data path and delivers committed changes in order and at least once to analytics, search, services, data lakes, and other downstream systems. Dedicated workers. Private connectivity. No connector stack for your team to deploy or babysit.

PostgreSQL today · 3 more sources planned · managed by Remac
Works with your stack
PostgreSQLMySQLsoonMongoDBsoonOraclesoonPostgreSQLMySQLsoonMongoDBsoonOraclesoon
Apache KafkaRabbitMQAmazon S3MinIOCloudflare R2PostgreSQLFileWebhooksQdrantRedisGoogle Cloud StorageAzure BlobDynamoDBDatabrickssoonMySQLsoonElasticsearchsoonSnowflakesoonBigQuerysoonPineconesoonKinesissoonClickHousesoonGoogle Pub/SubsoonRedshiftsoonApache PulsarsoonApache KafkaRabbitMQAmazon S3MinIOCloudflare R2PostgreSQLFileWebhooksQdrantRedisGoogle Cloud StorageAzure BlobDynamoDBDatabrickssoonMySQLsoonElasticsearchsoonSnowflakesoonBigQuerysoonPineconesoonKinesissoonClickHousesoonGoogle Pub/SubsoonRedshiftsoonApache Pulsarsoon
The status quo

Moving data shouldn’t cost you a team.

Today the options are heavy, expensive, or fragile. Here’s what data teams are actually living with.

Dashboards a day behind

Nightly ETL makes dashboards work from yesterday’s data. Time-sensitive decisions arrive late.

A cluster to babysit

Debezium runs on the JVM. Kafka Connect mode also adds Kafka and Connect workers. Other modes remove some infrastructure, but teams still operate a Java-based CDC path.

Bills that punish growth

Some managed connectors use row-based or usage-based pricing. Costs then grow with data activity and volume.

Several paths for one change

PostgreSQL replication moves data between PostgreSQL systems. Teams still need other paths when the same data must reach brokers, object stores, search, or another database engine.

Disasters with no undo

A scheduled backup cannot restore changes that happened after its recovery boundary. Continuous log archiving and replay create a finer recovery path.

Search that lies

Scheduled re-index jobs leave search behind the source between runs. Users can then miss items that already exist.

The collapse

Six capabilities. Two modes.

The industry often provides these functions through separate tools. They share one foundation. Logical functions use a decoded event path. Physical recovery preserves native log records and a compatible base snapshot.

CDC
Logical Replication
Heterogeneous Replication
Audit Logging
Point-in-Time Recovery
REDO / Replay
↓  six functions, one transaction-log foundation  ↓
MODE 01 · LOGICAL DECODING

Parse the log into ordered, row-level events. CDC, replication, heterogeneous sync, audit, and logical PITR share one decoded stream and one execution path.

postgres today · mysql · mongodb · oracle planned
MODE 02 · PHYSICAL ARCHIVING

Archive files or continuously stream raw log bytes for full-cluster recovery. The selected method determines the recovery point. No broker in the path.

file-based · streaming
What teams remac

One pipeline, a dozen jobs it retires.

Real-time analytics

Replace scheduled extraction with a continuous feed of selected committed changes.

Search index sync

Send selected changes to search as they commit instead of waiting for a full re-index.

Microservice sync

One source, many sinks, in order. Replace N polling integrations.

Audit & compliance

Preserve selected committed changes and their available source context in an append-oriented record.

Data lake ingestion

Write continuous JSONL to S3 and replace dedicated ingestion jobs where the pattern fits.

Vector DB sync

Tie embedding updates to selected committed changes instead of a full scheduled re-index.

Logical & heterogeneous replication

Converge selected row state in PostgreSQL or map it to a different database engine such as MySQL.

Point-in-time recovery

Restore a compatible full cluster from physical records, or reconstruct selected logical state from archived events.

REDO & event replay

Replay archived normalized events through configured sinks to rebuild, repair, test, or reconstruct declared state.

2
processing modes
6
transaction log capabilities
7
contract guarantees
74+
failure modes cataloged

“Every line of production code exists to move data from point A to point B without silent loss, broken order, or lost transaction context.”

Remac Engineering

“Trust comes from reliability. Reliability comes from testing. A rigorous test suite is not overhead. It is part of the product.”

Remac Engineering

“We handle other people’s data. A pipeline that silently drops events is worse than no pipeline because the operator believes replication is happening.”

Remac Engineering

“Six capabilities. Two underlying modes. One managed engine. We built one system around the full value of the transaction log.”

Remac Engineering

“The question is not only whether the pipeline works in normal conditions. It is what the pipeline does when its parts fail.”

Remac Engineering

“Every line of production code exists to move data from point A to point B without silent loss, broken order, or lost transaction context.”

Remac Engineering

“Trust comes from reliability. Reliability comes from testing. A rigorous test suite is not overhead. It is part of the product.”

Remac Engineering

“We handle other people’s data. A pipeline that silently drops events is worse than no pipeline because the operator believes replication is happening.”

Remac Engineering

“Six capabilities. Two underlying modes. One managed engine. We built one system around the full value of the transaction log.”

Remac Engineering

“The question is not only whether the pipeline works in normal conditions. It is what the pipeline does when its parts fail.”

Remac Engineering

One processor. Six transaction log functions.

Move committed data through one controlled path, with explicit ordering, delivery, and recovery boundaries.

Read the fundamentals