Workflows Working

Chain tools into one repeatable pass

Upload once, run every step in order, download the result. Workflows use the same tool definitions and the same processing contract as the rest of the site, so a chain behaves exactly as those tools do on their own pages.

Workflows run. Build a chain, set each step's options with the gear icon, add files, and each step feeds the next — a workflow of browser-side tools never leaves your device. Saved workflows are stored in this browser, because a workflow is a list of steps, not your data.

A repeated multi-step task — export a page as an image, remove its metadata, then compress it before it goes into a report — usually means opening several separate tools and downloading and re-uploading a file between each one. A workflow does the same steps in one pass: add each tool in order, adjust its settings once, and every file dropped in afterwards runs the whole chain without the manual handoffs. Templates below cover a few common chains as a starting point; anything can be rebuilt step by step, saved, and reused later without needing an account.

How workflows work

Every tool exposes one processing contract, so chaining is a matter of matching outputs to inputs.

One contract per tool

Each tool declares its input types, output types, settings and processing mode. The chain is checked against those before it runs, so a mismatch is reported up front rather than three slow steps in.

Runs where the steps run

A workflow made only of local tools executes entirely in your browser. One server step moves just that step to a worker — the badge on each card tells you which.

Saved in this browser

A saved workflow is a list of steps and settings, not your files, so it needs no account — and it stays in this browser. Passwords are never saved. Bulk and the API do not read saved workflows yet.