Connect
01
Connect your data systems and define how information should move between them.
Ingest
02
Move data in full, incrementally, or as a stream of changes, all through the same engine.
Deploy
03
Deploy Filament for local development, production infrastructure, or embedded in a service.
Extend
04
Use a growing list of prebuilt data storage and business tool connectors, or contribute your own.
impact
Transfer speed
in rows per seconds, higher is better
benchmark
Filament
2,597,498 ROWS\S
dlt
1,007,711 ROWS\S
PeerDB
718,368 ROWS\S
ingestr
488,233 ROWS\S
Sling
283,133 ROWS\S
Debezium
45,476 ROWS\S
Airbyte
28,698 ROWS\S
Rows moved per second over the whole load. NYC Taxi, 298,270,427 rows, Postgres to Postgres, median of three repetitions.
Wall time
in seconds, lower is better
Filament
115 SEC
ingestr
611 sec
Sling
1053 sec
Debezium
6559 sec
Airbyte
10,393 sec
Time from first row read to the destination reporting the full count.
CPU time
in seconds, lower is better
Filament
377 SEC
Sling
2,211 sec
PeerDB
611 sec
Debezium
611 sec
Airbyte
611 sec
Processor seconds burned across every container over the whole run.
Peak memory
in gigabytes, lower is better
Filament
2.97 GB
ingestr
4.83 GB
Airbyte
12.82 GB
Debezium
15.70 GB
dlt
45.54 GB
The highest memory a single container held at any point in the run.
Last updated on August 31, 2026
Why we built Filament
Filament is an open-source data movement engine built by data engineers, for data engineers. It carries information from wherever your data lives to wherever you need it, quickly and correctly, with the care of something meant to last and to hold up at enormous scale. It is Apache 2.0, it runs anywhere, and it is yours.
We built it because every good thing we have ever wanted to build sat downstream of the same unglamorous problem, getting the data in, and the tools for that were slow, expensive, or priced in a way that punished you at exactly the moment you needed them most. We have spent most of our careers on that problem, and we think you deserve infrastructure you can own. Filament is the first open-source piece of the context management system we are building at Galaxy.
What follows is the story of why it exists, what we believe, and where we are taking it.
Onwards,
Intergalactic Data Labs, Inc.
Ready to move data?
We'd love to help.
We'll integrate connectors you need.
We'll show you what you'd save, and how much faster it runs.
We'll walk through exactly what it takes to move from another platform.
And if you're thinking beyond data movement, we'd love to help too.
Prefer email? hello@getgalaxy.io
Learn more about Filament
resources
Quickstart
From zero to a running pipeline. Install the engine, connect a source, and move your first rows.
Read
Connectors
Every source and destination Filament speaks to out of the box, and how each one behaves.
Read
Deployment
Recipes for local development, standalone production setups, and embedding in an existing service.
Read
Contribute
Add support for a new system and get a consistent experience across every integration.
Read
galaxy-io/tempo
Keyboard-driven terminal client for inspecting and controlling Temporal workflows.
GitHub
galaxy-io/gnat
Keyboard-driven terminal client for managing and monitoring NATS JetStream.
GitHub
galaxy-io/sparktop
Live per-node GPU, host, and vLLM stats for DGX Spark clusters, in your terminal.
GitHub
used by teams at
Company
Talk to the team
Copyright © 2026 Galaxy. All rights reserved.