01 Source
Any source
Connect data wherever it lives. Credentials stay in the secret store.
ETL Pack · no-code · managed schema
No-code ETL Pack — trusted pipelines that move data wherever it needs to go, with breaking changes and schema drift blocked, not hoped away.
Built by operators with 8+ years running production warehouses and pipelines — now productizing the sync path we kept rebuilding. About →
01 Source
Connect data wherever it lives. Credentials stay in the secret store.
02 Detect
Catch adds, drops, and type changes before the warehouse absorbs them.
03 Sync
Watermark + upsert/merge on primary key. Full refresh when you choose.
04 Land
Land in the systems you already own — always yours.
05 Operate
Schedule, history, fail alerts, and one-click historical reload.
Freshness targethourly default
Drift blocked (breaking)drops / renames
Backfill windowdate / ID range
What we ownthe hard parts
What we refusesilent breakage
Most ETL pain is quiet: a column vanishes, a type widens, a rename ships green. Pack response is visual and deliberate — not a midnight Slack scroll.
Detected and held for explicit accept before downstream models assume it.
Flagged as risk. You choose widen / remap / pause — warehouse does not guess.
Blocked by default. Protects joins and dashboards that still need the field.
Green jobs, broken dashboards — how drift hides in type coercion.
BreakingRenames, drops, and type narrows that ship green and break BI later.
BackfillReload history without shredding watermarks or duplicating facts.
ChecklistKeys, types, and load modes that survive the first production week.
CDCWhen log-based CDC is overkill for warehouse correctness.
PricingWhen usage-based ELT bills spike — and when a flat pack is saner.
MARHidden row and credit costs that show up after the pilot looks cheap.
Answers
Yes — open the guided playground at 91crew.com/demo/ (no signup). It walks sync, watermark tips, schema drift blocked, and one-click recovery. Separate from the Custom Platform control-plane demo at demo.html.
Yes. New columns are held for explicit accept, type changes are flagged, and breaking drops are blocked by default — see drift controls or the silent drift guide.
Any source to any destination — databases like Postgres, MySQL, and Mongo, plus warehouses and systems you already run. One example path: Postgres → Snowflake checklist.
Often yes for warehouse correctness. When log-based CDC is overkill, read why DIY Debezium hurts.
No. ETL Pack is productized sync. For a Custom Data Platform in your infrastructure, see the Custom Platform.
Other offering
ETL Pack is productized sync. For a Custom Data Platform shaped to your stack and deployed where you run, see the Custom path.
Tell us your sources and destinations. We reply from hello@91crew.com with early access notes. Same crew that has run production data platforms for 8+ years. Need a full platform in your own infrastructure instead? Custom Platform →