WebToSparkThe quick answer
Choose Markdown when you want to read and annotate sources. Choose JSON when you need structured documents and nested replies. Choose CSV when you want to filter records in a spreadsheet. File format changes the representation, not how much of the page was loaded.
| Format | Best starting point | What to check |
|---|---|---|
| Markdown | Reading, notes and source material for your chosen AI tool | Readable structure and source context; not every structured field is rendered. |
| JSON | Programs, nested comment trees and record processing | Document schema, optional fields and batch envelope. |
| CSV | Excel, filtering and comparing rows | Record types, identifiers and JSON text in nested columns. |
JSON vs CSV: trees and rows
JSON can represent objects, arrays, numbers, strings, booleans and null. In WebToSpark it can keep a comments array with nested replies. CSV represents a table: the header names columns, and each subsequent record supplies cells. Nested structures need an explicit mapping to rows or cell text.
WebToSpark CSV uses record_type to distinguish document, primary, item and comment rows. For comments, record_id, parent_id and depth help relate rows. Columns such as metrics_json and links_json contain JSON text rather than separate spreadsheet columns. Empty fields can mean the page or record type did not provide that value.
Is this a JSON to CSV converter?
WebToSpark exports extracted page records directly to either format. It does not offer arbitrary JSON upload and conversion, and renaming .json to .csv will not transform the contents. For a nested JSON file, a separate conversion tool needs rules for which arrays become rows and which fields become columns.
If the source is already saved, request the missing format from the extension. This makes another representation of the available extracted source; it does not reconstruct comments that were never captured.
A practical research workflow
The website homepage offers fictional example downloads so you can compare the three formats. They illustrate format differences and are not a full production export schema. Free includes all three ordinary export formats within the local allowance.
- Open the source and load the material you need. Review the extension preview and warnings.
- Save JSON if downstream processing needs the structure, or CSV if your first task is filtering records.
- Use Markdown for reading or prepare a local Pro research packet when that capability is available to you.
- Check the output against the source. Keep capture context alongside any conclusions.