TASTEETL docs
TASTEETL packs a command-line tool (tasteetl) and a browser IDE into one folder. Name sources and targets as datasets, wire columns on the mapping canvas with lookups, filters and expressions, decide where and how a job loads, and order and schedule jobs in workflows. Check, trace and dry-run show real values without writing anything, so nothing is built blind. Runs leave a run log, lineage and rejected rows; post-load verification, quality rules and scores, schema drift detection, failure alerts and daily summaries follow.
Installation
We don't provide ad hoc download links until verified release files and URLs are registered.
What to check now
- Source version
- TBD
- Supported platforms
- Verifying release targets
- Release status
- Coming Soon
Quick start
- 1
Get the installer with a verified code signature and SHA-256 from the Downloads page and unpack it into one folder.
- 2
Launch tasteetl-ide and a browser opens. Create a project and register connections as env:name.
- 3
Build datasets and mappings on the canvas, confirm the real values with a check, then run the job.
Key features
Browser IDE
Explorer, tabs, canvas and a drawer (properties, preview, expression evaluator, messages, output) on one screen, in light and dark themes.
Mapping canvas
Drag to wire columns, attach lookups, add filters and expressions, and decide outputs and rejects. Node positions are saved next to the file.
Checks with real values
Check, trace and dry-run push a few source rows through and show the real value in every node, without writing anything.
Workflows
Chain jobs, set a condition on each link (success · failure · done), and choose parallel runs, retries and locks.
Schedules and the scheduler
Run daily, hourly or weekly, at intervals, when a file arrives, when a signal comes, or after upstream jobs. Spot overlaps on the calendar.
Four read modes
Pick full reads, incremental (watermark), snapshot comparison or the transaction log (PostgreSQL wal · MySQL binlog · Oracle redo · SQL Server CDC) per source.
Twelve sources and targets
Reads and writes PostgreSQL, MySQL, MariaDB, SQLite, Oracle, SQL Server, Db2, Altibase, CSV, Excel, Parquet and HTTP.
Post-load verification
Reconciles source and target by key, then records and reports missing, extra and mismatched rows.
Quality rules and scores
Put rules on a dataset and every run adds to its score; the run stops when the score falls below the threshold.
Schema drift detection
When a source table's columns change, it stops, or warns and carries on.
Rejected rows
Failed, filtered, not-found and blocked rows are kept in a file per kind; fix the values on screen and load them again.
Run log and lineage
Each run records rows read, written and failed, plus per-step counts, and two runs can be compared. Lineage shows where data came from and where it went.
Alerts and daily summaries
Failures are reported by email, webhook or command, and a summary goes out every day.
Releases and operations
Pin a version, open it run-only on the operations machine, and roll it back. Diagnostic bundles carry only definitions and history.
People, roles and API
Developer, operator and remote-support roles. Once people are registered, screens appear only after sign-in, and tokens let outside systems call the API.
Troubleshooting
I can't see a download button.
No public release files have been registered yet. Only ready files appear on the downloads page.
Can I check which operating systems are supported?
The docs only list environments verified for each product. “Being verified” means support isn't confirmed yet.
Where can I check the current feature scope?
The key features in these docs and the product page are written based on the current project source.