The workflow was never the hard part

You can build a good workflow in an afternoon. Right visual references. Compose the scene. Generate several PDP variants.

It works on one product.

Then the catalog arrives.

The workflow still works. But now you have to run it again and again, product by product. And somewhere around product forty, the thing that was supposed to save time has become another manual process

This is the part nobody demos. Building the workflow is the easy part to show. Running it at scale, checking every output, fixing the ones that missed, and keeping the same quality from product 1 to product 300. That is where teams lose time.

That gap is what Batch is for.

‍

Meet Batch

Batch is the production layer of Pletor. It takes a workflow you've already built and runs it across every input you have, in a single pass.

‍

‍

Bring a folder of packshots, an Airtable, a Shopify export, or a PIM file. Each row runs through the same workflow. Same logic. Same brand rules. Same output standard.

One workflow. Hundreds of products.

‍

Then you review everything in the same place.

Keep the results that worked. Fix or re-run the ones that need changes. QA stops being the wall at the end and becomes part of the run:

‍

And because the Pletor MCP makes your workflows callable, an agent can prepare the whole Batch for you. "Prepare a Catalog Ads Batch from my Q2 top performers, one row per SKU." The agent assembles the rows from data you already have; Batch runs them. MCP made workflows callable. Batch makes them runnable at volume.

Read more about MCP here.

‍

The point

A workflow you encode once should run on one input or five hundred without losing quality. Produce and review, at scale, in one place.

Batch is open in early access. Bring your largest input set and run it.

‍

Access the documentation here.