Now in beta

Stop stitching
tools together.
Run your data pipelines in Kusi.

Configure syncs, run jobs, and monitor results in one workflow. Start in the UI or define your workflows in Git. Choose shared or eligible dedicated execution.

Pay for usage. No subscription required.

Explore the product
ONE WORKFLOW. TWO WAYS TO START.

Start in the UI. Or start in Git.

Choose how you configure it.
Keep execution and monitoring together.

Same workflow. Your choice.
Configure a syncInteractive example
01 / CONFIGURE

Timezone: UTC

02 / YOUR WORKFLOW

Nightly orders sync

Existing tasks in this job

load.orders
load.customers
Also available as a YAML definition
Explore the same job through either view. Local example; nothing is saved or executed.

Run the work.

Run on shared workers, or use an operator-enabled dedicated compute profile for eligible workloads.

Explore execution

See what happened.

Follow runs, destination freshness, row volumes, and reported quality issues.

Explore observability

Pay for usage.

Buy credits for usage. Dedicated capacity is metered by lease time. No monthly subscription required.

How pricing works
CREDITS, WITH CONTEXT

Know what each run uses.

Follow credit consumption by job run. Shared runs use standard usage credits; dedicated capacity is metered by lease time.

Explore usage & pricing
01

Run-level visibility

Review credits alongside run history. Failed shared runs have no standard run charge; a ready dedicated lease can still accrue cost.

02

Server options affect cost Limited rollout

Eligible organizations can choose a named compute profile. Its hourly rate and lease lifetime determine the maximum credit reservation; unused credits return at settlement.

03

Estimate before you run Planned

We’re planning credit estimates for each job, informed by workload, past runs, and compute configuration, so you can review expected usage before starting.

LATEST RELEASE · v0.20.0

From loaded data to models.

Build dbt projects in your warehouse after a job succeeds, on a schedule, or on demand. Dedicated compute remains available for eligible workloads.

What's new in v0.20.0
01

Guided dbt setup

Check a Git-backed project before saving it, then inspect model and test results for each build.

02

Choose execution

Run transformations on shared workers or use a granted dedicated compute profile.

03

Alerts for builds

Get notified when a transformation build or its dbt tests fail.

Bring your next pipeline.

Start with one real workflow during the beta.

Get early access