Pipeline Passports
Issue a shareable trust warrant when checks pass. Auto-revoke on silent failure or KPI drift — the artifact execs and auditors can actually open.
DataXPipe is a pipeline contract with a live trust runtime — Passports, consumer contracts, KPI drift checks, and PR blast radius so green jobs and trustworthy data are the same event.
Issue a shareable trust warrant when checks pass. Auto-revoke on silent failure or KPI drift — the artifact execs and auditors can actually open.
Diff a pipeline spec in CI and see downstream datasets, consumer contracts, and passports at risk before merge — commented on the pull request.
Producers write the pipe. Consumers write the promise — max staleness, row floors, KPI checks — with compatibility status against live producers.
Catch wrong-but-green failures: kpi_baseline, row_count_drift, and distribution_drift against rolling baselines — not just freshness.
Register pipelines, datasets, and connections in one contract. Import dbt or Airflow — don't start from blank YAML.
Freshness and quality checks tied to runs. Alert when orchestrators succeed but data is stale — before stakeholders open the dashboard.
Issue a shareable trust warrant when checks pass. Auto-revoke on silent failure or KPI drift so execs and auditors never open a stale link.
Each passport captures SLA expiry, owner, lineage hash, and check summary — the artifact your board pack can link to.
Diff a proposed pipeline spec against the catalog and walk lineage multi-hop. See downstream datasets, consumer contracts, and passports at risk before merge.
GitHub Actions examples post the markdown report as a PR comment.
Dashboard and BI owners declare max staleness and row floors. Compatibility status compares those promises to producer freshness SLAs and active passports.
Broken contracts appear in PR blast-radius reports so merges don't wake Monday's board pack by surprise.
Wrong but green: freshness can pass while a metric drifts. kpi_baseline, row_count_drift, and distribution_drift compare against rolling baselines from prior passing runs.
Failures revoke passports and feed AI explanations that cite the last trusted warrant.
Freshness evaluation honors max_delay_minutes and threshold_hours server-side — status is derived, not caller-supplied.
Alert when orchestrators succeed but marts are stale, with lineage blast radius on the failure explanation.
Vision
Not another catalog to configure — a platform that imports what you already run, issues Pipeline Passports, previews blast radius in CI, and keeps producers honest against consumer contracts. One contract replaces scattered Airflow glue, quality scripts, and wiki docs.
dbt manifests, Airflow DAGs, and warehouse discovery — not greenfield YAML homework.
Freshness, KPI drift, lineage blast radius, and shareable Pipeline Passports.
Producer specs + consumer promises with PR impact analysis before merge.
Roadmap
We ship in the open. Early access users shape what lands next.
Available today
Silent failures, passports & PR blast radius
Import dbt/Airflow, catch green-but-stale failures, issue Pipeline Passports, and preview impact in CI.
Freshness and volume checks that fail when Airflow is green but data is stale.
Shareable trust warrants with SLA expiry, auto-issue on pass, revoke on fail.
POST /specs/impact + GitHub Action comments for downstream, contracts, passports.
Consumers declare staleness/row expectations; compatibility checks vs producers.
kpi_baseline, row_count_drift, and distribution_drift — wrong but green catches.
Bootstrap lineage and default checks from manifest.json or DAG source.
Stale demo → explain → passport → register in one session.
Q3 2026
Deeper runtime & authoring
AI-assisted spec drafting and richer hosted API surfaces for Day 2 teams.
api.dataxpipe.com — register pipelines and query runs from day one.
Failed checks routed to email and Slack incoming webhooks in Settings.
Click any dataset — see upstream/downstream blast radius visually.
Describe a pipeline in plain English; get a reviewable YAML contract.
Validate + impact on PR; register from protected branches.
2026 & beyond
AI control plane at scale
From assisted authoring to autonomous suggestions — always with human review before production.
Connect Snowflake/BigQuery — suggest contracts, owners, and freshness SLAs.
Suggested backfills, reruns, and owner escalations when checks fail.
Vetted specs for common warehouse and orchestration patterns.
SAML/OIDC, immutable audit logs, SOC 2 readiness.
Shape the roadmap
Request a workspace, tell us your stack (Airflow, dbt, Snowflake, BigQuery), and get priority when self-serve API access opens.
Spin up the API locally, generate the example orders_sync pipeline, and register it in under five minutes.