Workflow guides
Workflow guides show how Blocks work together to finish a browser task. Start with the smallest guide that matches your job, then add loops, variables, exports, and Cloud Runs after the basic path works.
Pick the smallest matching pattern
Most scraping and browser automation jobs fit one of these patterns:
- Open one page and extract one value.
- Extract repeated rows from a list, table, or search result.
- Click into detail pages and enrich each row.
- Fill a form, submit it, and read the result page.
- Export rows to a file, spreadsheet, email, API, or storage destination.
Start with one page and one output. A Flow that works on one page is easier to expand than a large Flow that fails in three places at once.
Build guides around visible results
A useful guide should end with something you can inspect: rows, text, a saved file, a log entry, or data sent to another tool. If the output is wrong, trace it back to the first Block that created or changed it.
For Block-specific settings, use the All Blocks reference. Guides explain how Blocks work together; Block pages explain the exact settings and outputs of one Block.
Add variables only when they remove repetition
Use variables when a later Block needs a value from an earlier Block. Do not add variables just to make a small Flow look more advanced. A hardcoded URL or filename is fine while you are proving the workflow.
When you do use variables, write the saved syntax directly: {{variableName}}.
Move to Cloud Runs after the Flow is proven
Use local runs while building and fixing. Move to Cloud Runs when the Flow has already produced the right output and you want hosted execution. If a Cloud Run behaves differently from a local run, compare the page state and the earliest Block that diverged.