Give your data shape. Visual ELT pipelines with CDC included
Visual extract, transform and load pipelines — with CDC included from the entry plan and an HTTP connector that AI wires up for you. The modern data stack, at a fraction of the cost of legacy tools.
The whole pipeline, under control.
From the source database to the destination — model, schedule and observe every step without leaving the canvas. Warehouse destinations (Snowflake, BigQuery) coming soon.
Visual pipelines
Assemble the extract → transform → load flow by dragging steps. Inspect the DAG and read each step's full execution history.
Source connectors
PostgreSQL, MySQL, SQL Server, Oracle and MongoDB with incremental, full and CDC sync — plus a generic HTTP connector that reads your API spec.
Transformations
Per-column conversion rules — cast, trim, regex, value maps, dates. Multi-table incremental with upsert by primary key.
Scheduling
Cron runs from every 5 minutes or trigger manually. Track success rate, duration and rows processed on every execution.
Multi-workspace
Per-customer isolation with its own subdomain. Sign in with Google or password — your data never crosses tenants.
Production-ready
Run observability, automatic retries and alerting. When something fails, you see the exact step, the error and the watermark.
Three steps from source to insight.
Connect the source
Point Januss at a database or API. We introspect the schema and list every table you can pull — pick what you need.
Model the transformation
Map columns, set conversion rules and write SQL transforms. Choose full, incremental or CDC sync per table.
Schedule and monitor
Set a cron schedule and watch it run. Every execution logs status, duration and rows — failures surface the exact step.
The HTTP connector that reads your API for you.
Drop in an OpenAPI or Swagger spec and Januss understands your endpoints, pagination and schema automatically. AI-assisted auth figures out the handshake — tokens, refresh, headers — so you connect any REST API without writing glue code.
- Parses OpenAPI / Swagger into ready-to-use endpoints
- AI-assisted auth: OAuth, API keys, bearer tokens
- Auto-handles pagination and nested response mapping
Convert and validate every column — no surprises.
Set conversion rules per column and decide what happens when a value doesn't convert. Then validate the transformed result with SQL before it ever reaches the destination.
- Convert types per column: string → decimal and back, to integer and dates.
- Adjust text case sensitivity — upper, lower and capitalize.
- On a conversion error, choose: fail the run or write null.
- Validate the result with SQL — if the query flags a row, the load is blocked.
-- block the load if any invalid row exists SELECT count(*) AS invalid FROM customer_360 WHERE amount IS NULL OR qty < 0; -- expected: 0 · otherwise the run fails
Sync on-prem sources without opening your firewall.
The Januss proxy agent runs inside your network. Source extraction and transformations happen on your local server — raw data is never sent to the cloud. Only orchestration metadata travels, over an outbound connection.
- No inbound firewall rules — the agent only makes outbound connections.
- Extraction and transformation run on your local server.
- Sensitive data stays inside your network — the cloud only orchestrates.
- Built for companies that can't expose data to the cloud.