Find the broken table before your dashboard does.
The open-source operator console for data platforms: lineage, event-driven scheduling, incidents and data trust, in one graph across the tools your data runs through.
Google Cloud today. AWS and Azure next.
- Run log
Not found: raw.orders_v2 - Schema change deployed 05:58 by MR
!214 - Downstream schedule held, 2 checks paused
Reads the tools your data team already runs
Lineage you never had to draw.
Nistor reads BigQuery job audit logs and parses Dataform and MicroStrategy SQL to map every table, job and dashboard, down to the column. Pick any asset to see what feeds it, and what breaks if it does.
- Column-level lineage, each edge with a confidence score
- Blast radius for any change, before you merge it
- A shareable Resource Passport for every asset
Run when the data is ready, not when the clock says.
Event-driven rules start Dataform, Data Fusion, BigQuery and MicroStrategy jobs the moment their upstream data lands, with debounce, cost gates and a lab to replay them against past events first.
- Rules, alerts and contracts live in git
- An MR bot posts the blast radius of every merge request
- Alerts to Slack, Google Chat, email and webhooks
Planned: the same rules start AWS Glue jobs, Step Functions and Azure Data Factory pipelines.
From red to root cause on one screen.
Failures open incidents on their own, grouped by error. Nistor walks the graph to score the probable cause, and shows exactly what is impacted downstream, before anyone opens a ticket.
- Probable cause scored from runs, deploys and schema changes
- Hold downstream schedules so bad data stops spreading
- Every destructive operation gets a time-boxed, audited undo
- Dataform run stg_orders failedNot found: raw.orders_v2
- Incident opened, 14 errors groupedOwner data-platform notified on Slack
- 3 dashboards impacted, downstream schedule held
- Probable cause: schema change in MR !214 0.92
- Fix merged, run replayed, data fresh againHeld schedules released automatically
Promises your data consumers can check.
Contracts for freshness, volume, schema and quality, SLOs per health group, and a status page analysts can bookmark. Import and export ODCS contracts.
Built for every cloud. Starting with Google Cloud.
Nistor's core names no vendor: warehouses, object storage, message buses and secrets plug in through open-source contracts, and each cloud arrives as its own connector. Google Cloud is the first, and the deepest. AWS and Azure are next, so one graph holds the lineage, schedules and spend of every cloud you run.
| Capability |  Google CloudToday |  AWSNext |  AzureNext |
|---|---|---|---|
| Lineage & assets | Available25 asset types, and column lineage from BigQuery audit logs | PlannedS3, Glue Data Catalog, Redshift and Athena | PlannedData Lake Storage, Synapse, Microsoft Fabric and Azure SQL |
| Scheduling | AvailableRules run Dataform, Data Fusion and BigQuery jobs | PlannedRules run Glue jobs, Step Functions and Amazon MWAA | PlannedRules run Data Factory and Fabric pipelines |
| Incidents & trust | AvailableFailed runs, missed freshness and broken contracts | PlannedFailed Glue and Step Functions runs, late data | PlannedFailed Data Factory and Fabric runs, late data |
| FinOpsEnterprise | AvailableBigQuery job cost by table, team and user | PlannedCost and Usage Reports, Redshift and Athena query cost | PlannedCost Management exports, Synapse and Fabric capacity |
| Data qualityEnterprise | AvailableChecks on BigQuery and Cloud Storage | PlannedChecks on Redshift, Athena and S3 | PlannedChecks on Synapse, Fabric and Data Lake Storage |
| Install Nistor | AvailableCloud Run, Cloud SQL and Pub/Sub, with the Terraform in the repo | Coming soonIn your own AWS account | Coming soonIn your own Azure subscription |
Tools that are not tied to a cloud connect today, wherever they run: MicroStrategy, GitLab, Google Workspace and ODCS contracts, plus Sifflet and Castor with Enterprise.
Running on AWS or Azure? Tell us what you run →The graph, everywhere you already are.
CLI
Validate and sync rules, sync alerts and contracts, in CI, with JSON reports.
MCP server
26 read-only tools for Claude, Cursor and Gemini CLI, 15 of them open source.
VS Code
Hover a BigQuery table in a .sql file for its open incidents, upstream and downstream.
Chrome extension
Health, lineage and incidents inside the GCP console.
Nistor's view of the console page you have open.
Open a table, a service or a topic in the Google Cloud console. Nistor Inspector reads the resource from the tab's address and shows what Nistor knows about it, without leaving the page.
- Incidents on the resource and upstream of it, by severity
- Quality checks, 30 days of cost and lineage both ways
- Recent runs, their success rate and the rules that schedule them
- ⌘K to search nodes, incidents and dashboards
It knows BigQuery tables and datasets, Cloud Run services, Cloud Scheduler jobs, Cloud Storage buckets, Pub/Sub topics and subscriptions, and Dataform repositories.
1 active critical
90% pass rate 9/10 checks
$18.20↑ 4%
2 upstream · 3 downstream upstream incidents
93% success 2 rules
A badge that counts incidents
On a page it knows, the toolbar icon counts the incidents on the resource and upstream of it, or shows ! when they are all upstream.
Ask AI in the side panel
Ask about the open resource in Chrome's side panel and get answers with their sources. With the Enterprise AI module.
Reminders on your desktop
A reminder set on a node comes due as a desktop notification, and the badge counts the ones waiting.
The popup and the badge read from the FinOps module of Nistor Enterprise, and Ask AI from its AI module; sign-in and reminders need only the open-source core. Nistor Inspector is not in the Chrome Web Store yet: you build it from the source and load it unpacked in Chrome.
Yours to run. Apache 2.0.
Nistor runs in your own cloud project, today on Google Cloud: Cloud Run, Cloud SQL and Pub/Sub, deployed with the Terraform modules in the repo. Installing on AWS and Azure is coming soon. Try it on your laptop first.
- Get the source: it opens on GitLab soon.
cd nistor && docker compose up- Open localhost:8080: a seeded 35-day estate with 216 runs and 6 incidents. No cloud account needed.
The deployment guide comes with the docs, soon.