Skip to main content
Filament moves data from sources to sinks with full, incremental, and CDC replication. Every batch is integrity checked on both the read and write side, and interrupted runs resume from checkpoints. It is written in Go, and its major pieces are pluggable: connectors, the data store, and the event bus are interfaces with swappable implementations.

​
How it’s put together

Filament is three components over a data store and an event bus.
  • server serves the API and web UI and owns connections, pipelines, and run requests.
  • control-plane dispatches requested runs and tracks run state.
  • worker executes a single run from extraction to verified write.
Read more about how these fit together in Architecture.

​
Where to go next

Quickstart

Get infra, the control plane, API server, and UI running locally.

Architecture

Understand the server / control-plane / worker split.

Connectors

See what Filament can read from and write to today.

Deploy on Kubernetes

Install the Helm chart against your own Postgres and NATS.
Filament is pre-1.0. APIs — including the interfaces documented here — may still change.