Data pipelines for teams on Google Cloud
Drag sources, transforms and targets onto a canvas. PipeKeel writes the SQL and runs it — inside your own Google Cloud project, or fully managed by us. No code to maintain, no data copied out of your cloud unless you choose to.
Now in beta · 14-day trial of the Team plan, then a free plan. No credit card.
Choose per pipeline. The canvas, blocks, schedules and alerts are the same either way.
Pipelines run as BigQuery jobs in your project, as a service account you control. Your rows never pass through PipeKeel — we send the SQL and read job status. Databases and files outside BigQuery are moved by a Runner that also runs in your project.
No cloud setup at all: PipeKeel runs each task on DuckDB or BigQuery in your organization's region — United States, Canada or India — with storage reserved for your organization.
Add BigQuery, Cloud Storage, Google Sheets, MySQL, PostgreSQL, SQL Server, Oracle, SFTP, files over HTTPS, any REST API, Pub/Sub or Kafka. Test the connection, browse tables and columns.
Drag blocks onto the canvas and connect them. Problems show as you go, preview rows at any step, and see the exact SQL PipeKeel will run.
Save a version, activate it, and run it on a schedule, from a webhook or after another pipeline. Chain pipelines into workflows; get alerts by email or Slack.
Joins, aggregates, window calculations, pivots, deduplication, JSON flattening and unnesting, XML parsing, masking personal data — custom SQL when you need it, and 39 templates to start from.
Read only new or changed rows, upsert into targets, or keep full history with slowly changing dimensions — tested on BigQuery and DuckDB.
Not null, unique, accepted values, ranges, patterns, freshness and row counts — fail the run or just warn.
Run pipelines in order, call HTTP endpoints, send email or Slack messages, branch on conditions and fill in values from earlier steps.
Build in dev, promote to UAT and production with a diff of every change. Protected environments accept changes by promotion only, with approvals on Business.
Dates, dataset names and API tokens per environment — bound safely as query parameters, secrets write-only.
Roles, teams and per-pipeline permissions, required two-factor authentication, IP allow-lists, and a tamper-evident audit log.
Estimate a run before it starts, cap the bytes each query may scan, and see BigQuery bytes and Express credits on every run.
Real time from Pub/Sub and Kafka — filters, derived columns and masking applied as records arrive, in your own project for Sovereign — or micro-batches with every block. Publish results back to topics.
Append, overwrite, merge (upsert) or SCD2 — chosen per target; APIs and topics take each row as JSON.
Security isn't a feature we added — it decides where things run.
How we protect your dataThen keep a free plan, or choose a paid one from $49 / ₹1,999 a month.
