Maxun: Record Pagination and Export Data without Code
Configure a Maxun Recorder list, verify field alignment and next-page actions, rerun the saved robot, and inspect exported records for omissions and duplicates.
A no-code recorder can turn a small directory into a repeatable extraction task. A correct preview, however, does not establish a correct saved run, and forty exported rows do not prove coverage of two distinct pages.
This tutorial collects displayed titles, displayed prices, and stock text from Books to Scrape. It configures next-page navigation and a forty-row cap, then exports and checks the result. Detail-only UPCs and descriptions are outside this task.
Prepare the input addresses
Locate the following pages in your browser, confirm their contents, and copy their full addresses. Obtain service and repository addresses from account integration settings, official setup instructions, or the repository clone menu. Do not derive real addresses from illustrative output.
BOOKS_LIST_URL: the first Books to Scrape catalogue-page address.
Replace each <VARIABLE_NAME> in JSON or task parameters with the copied value before submission; JSON does not expand environment variables. Any example.16yun.cn source in an output illustration shows field structure, not a demo service deployed for this article.
Enter a working workspace
Use Go To App on the official project page, or your team's existing instance. Sign in and confirm permission and capacity to create and run Recorder tasks. Hosted functionality and charges depend on the account.
This path uses the workspace as its environment; it does not require installing a separate database or model. Self-hosting is a separate deployment task governed by the selected release's instructions.
Create an Extract task in Recorder Mode, not AI Mode. The list-and-pagination workflow does not require a model key. A remote browser must be able to reach the target independently of your desktop.
Establish two-page checks before recording
Open the first two catalog pages manually and note two recognizable books and their displayed prices on each. The site uses demonstration prices.
| Column | Meaning | Limitation |
|---|---|---|
title_display | Visible card title | May be truncated |
price_display | Price text from the same card | Not yet normalized money |
availability_display | Stock wording | Not a verified stock quantity |
detail_url | Link address if supported by capture | A source link does not mean the detail was visited |
If the capture returns only link text, omit the URL column rather than relabeling a title as an address. Preserve entry URL and task name as batch context.
Record the repeated list
Enter this address and choose Start recording:
<BOOKS_LIST_URL>Confirm the recorder shows the catalog, then use Capture List on a repeated product card.
One card should become one row. An entire page in one cell means excessive scope. One isolated book means the repeated list was not identified. Mixed titles and prices indicate broken row association.
Fix this before adding pagination.
Select and inspect fields
Capture title, price, and stock within the same repeated object. Name columns according to their meaning. After each addition, compare the first and a middle product rather than assuming every row follows the first.
Use the confirmation action appropriate to the current capture stage. Do not end the task merely because a button says Confirm.
A visible title may contain an ellipsis. That is not an export encoding error. If full names are required, determine whether the capture supports the link's title attribute. Otherwise use a detail task or another suitable method rather than inventing missing text.
Configure next-page behavior
Choose Click Next and inspect the detected next control. Use Choose Different Element if it selected navigation or a page number instead.
Set the custom maximum row count to forty. At twenty books per page this bounds a small teaching run, but it is a row cap, not a universal two-page guarantee. Changed page sizes, deduplication, or missed rows affect that relationship.
Click Load More and scrolling describe different list behaviors. They should not replace a conventional next-page action simply because each can produce additional data.
Save and rerun independently
Finish, name the robot, and save it. A recording without a valid capture action is not a complete extraction definition.
Run the saved robot from the task list. This rebuilds its steps rather than relying on the recorder's current page state.
In the captured-list results, find your second-page samples, then your first-page samples. A total count alone could hide repeated collection of the first page.
Export and reconcile
Use Download options in the actual run result and choose JSON or CSV. See the official project for its result interface.
This is an illustrative CSV shape, not a downloaded run:
title_display,price_display,availability_display
A Light in the ...,£51.77,In stockRetain the original export and edit a separate copy. Select the intended encoding and text types in spreadsheet software to avoid automatic identifier or number conversion.
Check column meaning, missing values, source samples, and duplicates. If links are available, use them as stronger duplicate evidence. Equal truncated titles and prices establish only a possible duplicate.
Two books may share a truncated display title. A reduced row count can therefore signal identity loss rather than an insufficient row cap.
Troubleshoot
Recorder does not open: inspect account access, browser capacity, and remote reachability. Self-hosted browser/backend connectivity belongs to the environment stage.
Only one row previews: reselect the repeated card region rather than an isolated title.
Preview works but the saved run is empty: compare entry state, capture actions, readiness, and the current run's error. Recheck a small capture rather than trusting old previews.
Only page one appears: check the cap, selected next control, and actual content change. Require a known page-two sample before scaling.
Prices belong to other products: restore field scope within the repeated object and verify sample rows before export.
Pagination never stops: stop the task and inspect cycling, unchanged content, and bounds. Rerun with a small cap after correction.
Columns retain names but values change after a redesign: names are labels, not evidence that selectors remain correct. Recheck position and row association.
A repeatable small recorder task provides a useful foundation before adding login, categories, and detail navigation.
Need an enterprise proxy plan?
We can tailor architecture to your target domains, concurrency, and reliability goals.