Docs
Your first pipeline, start to finish.
Install the CLI, describe a pipeline in twelve lines of YAML and deploy it. Everything below works on the free plan.
01
Quickstart
Install nothing globally. The CLI runs through npx, asks for a source and a destination, and writes a sluice.yaml next to your code.
terminal
$ npx sluice init
? Source postgres
? Destination snowflake
✓ wrote sluice.yaml
02
Configure
A pipeline is one file. Name it, point it at a source, list the tables and pick a destination. Transforms run in order, and dedupe on id turns on exactly once delivery.
sluice.yaml
pipeline: orders_to_warehouse
source:
type: postgres
tables: [orders, refunds]
transform:
- dedupe: id
destination:
type: snowflake
schema: analytics
03
Deploy
Deploy from your laptop or from CI. Sluice validates the config, connects both ends, snapshots the tables and switches to streaming. Every branch can get its own preview pipeline.
terminal
$ sluice deploy orders_to_warehouse
✓ validated config (12 lines)
✓ connected postgres · wal
✓ exactly once enabled · key: id
● streaming
04
API
Everything the CLI does goes through the REST API, so you can create, pause and replay pipelines from your own tooling. Authenticate with a project token.
rest
$ curl $SLUICE_API/v1/pipelines \
-H “Authorization: Bearer $SLUICE_TOKEN” \
-d @sluice.yaml
05
SDKs
Typed clients wrap the API for the languages most teams use. Each one ships with the same retries and idempotency keys as the CLI.
install
# TypeScript
$ npm install @sluice/sdk
# Python
$ pip install sluice
# Go
$ go get sluice.dev/go
# Terraform
provider “sluice” {}