
Blazing-fast open-source data replication and CDC
Filament is an open-source (Apache 2.0) data replication and change data capture (CDC) engine built by Intergalactic Data Labs (Galaxy). It moves data between systems via full loads, incremental syncs, or streaming changes through a single engine, with native Postgres and MySQL CDC. The team reports 2.6M records per second in benchmarks against the NYC Taxi dataset and positions it as infrastructure data engineers can own rather than rent — the first open-source component of the context management system Galaxy is building.
Move data in full, incrementally, or as a stream of changes (CDC), all through the same engine, with pipelines defined by connecting sources and sinks.
Change data capture is supported natively for Postgres and MySQL sources.
Benchmarked at 2,597,498 rows/second moving 298M NYC Taxi rows in under 2 minutes (115 seconds wall time, 377 seconds CPU, 2.97 GiB peak memory), Postgres to Postgres.
A growing list of prebuilt data storage and business tool connectors, with documented guidance for building and contributing your own.
Per-pipeline observability with run counts, record counts, volume, and throughput, viewable across 1h/24h/7d windows and pivotable by pipeline.
Deploy for local development, standalone production setups, or embedded in an existing service, with published deployment recipes.
Fully open source under Apache 2.0, runs anywhere, and meters nothing.
Connect your data systems and define how information should move between them.
Move data in full, incrementally, or as a stream of changes through the same engine.
Deploy for local development, production infrastructure, or embedded in a service.
Use prebuilt connectors or contribute your own for a consistent experience across integrations.
Filament is open source under the Apache 2.0 license and, per the site, 'meters nothing' — the site invites users to 'get started for free.' No paid plans or tiers are listed on the page.
Filament is an open-source data replication and CDC engine that moves data from wherever it lives to wherever it's needed — in full, incrementally, or as a stream of changes — all through the same engine.
Yes. Filament is Apache 2.0 licensed, runs anywhere, and meters nothing. The maintainers describe it as infrastructure you can own.
Yes, with native CDC support for Postgres and MySQL sources.
Filament offers a growing list of prebuilt connectors for data storage and business tools. The page shows example pipelines such as Postgres to ClickHouse and a Slack sync; the full connector list is documented in the docs.
Filament can be deployed for local development, standalone production setups, or embedded in an existing service. Deployment guides and recipes are available in the documentation.
In public benchmarks on the NYC Taxi dataset (298M rows, Postgres to Postgres), Filament moved 2,597,498 rows per second, completed the load in 115 seconds, and used 2.97 GiB peak memory — 91x faster than Airbyte and with 15x less memory than dlt on identical hardware.
The site's FAQ lists data verification as a topic, but the specific verification mechanism is covered in the documentation rather than detailed on this page.
Filament is built by Intergalactic Data Labs, Inc. (Galaxy), by data engineers for data engineers, and is the first open-source piece of the context management system they are building.