The Scenario
You inherited the QA workbook. The previous engineer who built the test harness left three weeks ago, and the workbook still has 60 URLs in column A from last sprint — product pages, checkout flows, and a handful of dynamic cart routes. Your job this week is to verify none of them are throwing JavaScript errors before the release goes out Friday.
The bad version:
- You export the URL list to a CSV, write a Playwright script locally, and run it against Browserbase — which means configuring your API key, setting session parameters, and remembering the exact endpoint format.
- The logs come back as nested JSON. You write another script to extract the error messages, match them back to the original URLs, and build a flat table.
- You paste 60 rows into the workbook, discover three columns are misaligned because two URLs had no errors and skipped a field, and spend 20 minutes fixing the offset.
The release is in two days. Debugging your data pipeline is not the same as debugging the product. You need the error flags in the workbook, not a working ETL script.
The Easy Way: One Prompt in SheetXAI
SheetXAI is an AI agent that lives inside your Excel workbook. It reads the data in front of you and talks to Browserbase Tool on your behalf — creating sessions, fetching logs, and writing results back into your columns. No scripts. No JSON parsing. You describe what you want.
Create a Browserbase browser session for each URL in column A of my "QA URLs" worksheet, retrieve the session logs, scan for JavaScript errors, and write "ERROR" or "PASS" into column B and the first error message (if any) into column C.
What You Get
After running that prompt:
- Column B fills with "ERROR" or "PASS" for every URL row.
- Column C contains the first JavaScript error message for any errored session, blank otherwise.
- Rows where the session itself failed to create (invalid URL, network timeout) get flagged separately with a note like "SESSION_FAILED" so they're easy to filter.
- The original column A URLs are untouched.
What If the Data Is Not Quite Ready
The URL column has mixed formats — some have trailing slashes, some don't, some are HTTP instead of HTTPS
Normalize every URL in column A before creating sessions: strip trailing slashes, upgrade http:// to https://, and remove any query parameters. Then run a Browserbase session for each normalized URL, write the cleaned URL into column D, and write the session status into column E.
Some rows have already been tested — column B already has a value
Skip any row in column A where column B already contains "ERROR" or "PASS". For all remaining rows, create a Browserbase session, retrieve the logs, and fill in column B and column C with the result.
The URLs span two worksheets — "Sprint A" and "Sprint B"
Pull the URLs from column A in both the "Sprint A" and "Sprint B" worksheets, run a Browserbase session for each, and write the error status and first error message back into columns B and C of the respective worksheet.
The URLs are mixed — some need authentication context, some don't; flag them separately, then run sessions and log all results in one pass
Look at column C in my "QA URLs" worksheet — if it says "auth required", note that in the output. For all URLs in column A, run a Browserbase session, retrieve the logs, and write the session status into column D, the first error message into column E, and whether the URL was flagged as auth-required into column F. Do this in one pass across all rows.
One prompt handles the conditional tagging and the session run together — no separate cleanup pass needed.
Try It
Get the 7-day free trial of SheetXAI and open your QA workbook with a column of URLs, then ask it to spin up Browserbase sessions and write back the error status for each. You can also check out how to export a usage report for cost forecasting or audit failed sessions for debugging.