Need deeper analysis? See the related editorial article for background and trade-offs. Use this guide for the practical steps.
Batch processing can turn hours of opening separate files into a short review, provided the tool keeps results separate, avoids unsafe archives and never silently changes a production source.
Use this checklist alongside the local checker. Every step is free and can be repeated without an account or daily quota. Only approve corrections that reflect the real page. Where Google eligibility matters, follow the linked official documentation rather than treating a local result as certification.
Step 1: Prepare a manageable batch
Collect relevant .json, .jsonld and .html files into a folder or ordinary ZIP and keep clear source filenames.
Look for: Sending hundreds of unrelated files through one operation makes the result hard to review and can exhaust browser memory.
Example: Group product templates separately from Organization and breadcrumb files so findings have useful context.
Keep in mind: An unlimited daily usage policy does not mean unlimited memory per individual browser operation.
Step 2: Preserve file identity
Validate each file independently and attach the source filename to diagnostics, reports and corrected downloads.
Look for: Every document can have different structured-data types and errors. Combining their source text into one blob produces misleading counts.
Example: Two files may each declare an Organization, but a malformed field in only one must remain traceable to that file.
Keep in mind: Do not rewrite file paths into locations outside the user’s chosen download area.
Step 3: Inspect archives before parsing
Reject unsafe paths, enforce a per-file decompressed size limit and verify CRC checksums when handling ZIP contents.
Look for: ZIP archives can contain path traversal names, encrypted entries or compressed content much larger than the ZIP itself.
Example: An archive containing ../private.txt is not a valid name to use for a corrected file download.
Keep in mind: Password-protected and unsupported compression methods should fail clearly rather than partially processing.
Step 4: Keep processing on the device
Read browser File objects, apply parsing in JavaScript and create downloadable Blob objects without sending file contents to a backend.
Look for: Local analysis is especially helpful for agencies handling prelaunch HTML that should not be transmitted to a new third-party validator.
Example: A staging product HTML file can be checked before deployment without disclosing its private page details.
Keep in mind: Hosting providers still see normal website requests; local processing is not complete online anonymity.
Step 5: Give each issue its own explanation
Show per-file errors, warnings, entity count and a detailed issue view with official references.
Look for: A single red batch score makes triage difficult. Teams need to know what is wrong in each file and how to repair it.
Example: Open the file with a Product price problem separately rather than hiding it under a total of twelve warnings.
Keep in mind: Do not replace specific diagnostics with a vague “batch failed” message.
Step 6: Export corrections without guessing
Create corrected files only when parsing and deterministic fixes succeed, then allow the user to inspect the result and download a ZIP.
Look for: Some syntax errors can be safely repaired, but factual omissions cannot be solved by generic invented content.
Example: A missing product price should remain a documented manual task rather than being replaced by a made-up figure.
Keep in mind: Never mutate the original uploaded files in place.
Final review checklist
Recheck syntax and warnings after editing, compare the corrected file against the original, export a report, and validate applicable features using official tools. If a result is unclear, keep the original data instead of guessing business facts.
Official sources: Google documentation and additional reference. For more detail, read the complete article.
Frequently asked questions
Answers are educational. FAQ markup does not guarantee Google FAQ rich results.
Does the ZIP batch upload files to a server?
No. The current batch workflow processes supported files in the browser.
Will one invalid file invalidate every other file?
No. Review each file and its own diagnostics separately.
