A webpage table. A usable spreadsheet. Less copying.

Start with the table you actually need

A supplier directory, a pricing comparison, a research shortlist: useful information often starts life as a table on a webpage. Before copying anything, decide which rows and columns belong in your output. A clean export begins with a clear scope.

Check whether the page offers its own CSV or spreadsheet download. A native export is often the quickest way to get complete data, especially when the table has pagination or hidden columns. If it does not, copying the visible cells can work well for a small table.

Do not assume the information you see is the entire dataset. A page might load more rows as you scroll or show only one page of results. Record the source URL and the part of the table you captured so you can revisit it later.

Extract the table with WebAct

The free website table extractor includes an example CSV and a ready-to-copy task. Install WebAct, verify your account and choose a ChatGPT browser connection. Activate the extension on the source tab and select the table as context.

Switch to Task mode and paste the extraction task. It asks for the original headers and values, a source URL column, and a clear count of exported rows. Start with a small table of up to 25 loaded rows. Review the proposed actions, run the task, and check the file before importing it.

Compare the exported row count with the rows you intended to capture. Inspect the first row, the last row, and a row with a comma or quotation mark in it. Missing or unreadable content should be reported, never filled in by guessing.

Check the file in your spreadsheet

Use your spreadsheet’s import flow when the data contains ZIP codes, product codes, or long numeric IDs. Treat these columns as text so leading zeros and large identifiers stay intact. A CSV carries values and delimiters, but does not preserve spreadsheet formatting or column types.

The extraction task asks WebAct to neutralize cells that could be interpreted as spreadsheet formulas and report changed values. Check those values before opening the file. CSV does not carry formatting or reliable type information.

If the next application needs JSON, ask for the exact field names and value types it expects. Check one complete record before using the result.

When copying becomes the job

For a one-off table, a native download may be enough. For repeated extraction, save a WebAct task with the columns and output format you need.

With WebAct active on the page, select the relevant region and ask: “Extract the visible names, companies, and roles into a CSV. Include the headers and check that every visible row is present.” Review the result against the source, then save the task if you want to reuse it.

Be explicit about “visible.” Hidden rows, other pages, and content the browser has not loaded should never be assumed to be included. A useful automation gives you a result you can check, with a clear explanation of its coverage.