CLOUDFLARE EDGEALLERMOI · ON-DEVICE GEMMAMRPIPE · DRY-RUN + ROLLBACKMANILA 14.6°N 121.0°EHUMAN IN THE LOOPEVERY CLAIM HAS A SOURCE
IN ACTIVE DEVELOPMENT · Web

Mr. Pipe

Move data between the platforms a business already runs on, without a spreadsheet and without a one-way door.

BEFORE ANYTHING IS WRITTEN IN ACTIVE DEVELOPMENT
Schema read from live orgnok
DRY RUNnAPPROVES EVERY MAP
BEFORE EVERY PUSH
01 · THE PROBLEM

Moving data between systems is a one-way door.

A business changes CRM, adds a project tracker, or moves reporting into a spreadsheet, and someone has to carry the records across. That work is usually done by hand, or by a script written once and never read again.

Field names do not match. Picklists do not match. A phone number is formatted six ways, a required field on the target has no source at all, and a currency lands in a text box. None of that is visible while it is happening. It becomes visible afterwards, in the target system, at scale.

Mr. Pipe makes the write the last step instead of the first. It reads both schemas, shows every field decision with a confidence on it, runs the whole thing against real records without writing anything, and only then asks.

02 · HOW IT WORKS

Five steps, one of them a person.

01 CONNECT

Connect both sides

Pick a source and a target, enter credentials, verify. A verified connection reports back who it is: the org name, the workspace, the account.n02

03 · THE GATE

The model is not trusted with the write.

Field mapping is a good use of a model. It is pattern matching across two vocabularies, and a model does it faster and more completely than a person reading two schemas side by side. What a model should not do is act on its own output.

So the confidence score is shown rather than used as a threshold to auto-apply. Low confidence is not hidden, it is sorted to the front. The reasoning is written out per row, so a decision can be disagreed with instead of merely accepted. Approval is one deliberate action over a reviewed set, not a click that happens per row while attention drifts.

And the last gate is mechanical rather than a matter of care: the dry run cannot write. It produces the preview and stops. The interface says so on the job itself, while the job is running.

Nothing written, preview only

04 · SCREENS

Everything on one surface.

WEB
Mapping review: an object-pair ribbon across the top, a stats bar, filters, and a table of field mappings each with a transform, a confidence bar and an approve action.
Mapping review. One row per field. The ribbon at the top is the object pairing and its confidence; the table below is where the decisions actually get made.
A completed dry run showing a before and after table of sample records, with changed values highlighted, and an approve and execute action.
Dry run complete. The whole job ran. Every changed value is marked, the transform count and validation errors are stated, and nothing was written.
Job history table listing past dry runs and live pushes with status, record counts, error counts, duration, and a rollback action.
History. Runs are kept with their record counts, errors and duration. A completed live push keeps its rollback available; a dry run has nothing to reverse.
05 · STACK

Under the hood.

STACK
Interface React and TypeScript, desktop-first, darknService

Connects to

CRM PlatformsnProject and Issue TrackersnSpreadsheetsnCloud StoragenDocument EditorsnCalendarsnEmailnAd PlatformsnVideo PlatformsnProspecting ToolsnTime TrackingnMarketing AutomationnEcommerce and FulfillmentnDocumentation Workspaces
PARTNER

Got a migrationnyou're dreading?

Mr. Pipe is in build. If you have data to move between the platforms above, tell Zedlav what and where.

sales@zedlav.ai →

Want one of
these built?

Zedlav builds client work the same way it builds its own products. Send what is slow, manual, or breaking.

sales@zedlav.ai →