The .zip file.
The way to send twelve monthly exports, a partitioned dataset, or a folder someone gave you without opening it first.
What it actually is
An archive is unpacked and each file inside is treated exactly as though you had uploaded it yourself. Formats that can be read are read, formats that cannot are named, and nothing is silently discarded.
This is the right way to send anything that is one dataset split across many files. Twelve monthly CSVs with identical headers are recombined into one table with the source file kept as a column, so a finding can still point at the month it came from. A partitioned Parquet directory is reassembled with its partition keys restored from the folder names.
It is also the honest way to send a folder you have not audited. The contents list comes back with what each file is and what will be done with it, which is frequently the first time anyone has looked.
At a glance
- Unpacked
- Every file inside is treated as its own source
- Split datasets
- Files sharing a schema are recombined, keeping the source name
- Partitioned exports
- Keys encoded in folder names are restored as columns
- Encrypted archives
- Reported, never silently skipped
- Read here as
- Extracted, then each member on its own terms
- Status
- Ready
- Fidelity
- Data extracted
- Largest file
- 250 MB
Unpacked, and each file inside is treated as though you had uploaded it yourself.
Encrypted archives cannot be opened, and are reported rather than silently skipped.
Upload one
The analysis runs before there is anything to pay for. You see what it found, and the evidence behind it, first.
No account needed to start. You only pay when you like what you see.
What goes wrong, and what is done about it
01
Twelve files that are almost the same schema
Monthly exports drift. A column is added in March and renamed in August, and stacking them naively produces a table where the same measure occupies three columns, each mostly empty.
Columns are matched across members by content as well as by name, and any schema change between files is reported with the month it happened in.
02
A nested archive nobody expected
Exports that bundle their own attachments produce zips inside zips, and a shallow unpack finds a single unreadable file where the data was.
Nested archives are unpacked to a bounded depth, and anything left unopened is named in the contents list rather than left as a gap.
Questions people ask
Does the archive count as one file against my limit?
No. What counts is what comes out of it, because that is what gets analysed. The contents are listed before anything is charged for, so an archive holding one spreadsheet costs exactly what uploading that spreadsheet on its own would have.
What about a password-protected archive?
It cannot be opened, and you are told so directly rather than discovering the gap later in the findings. Nothing is guessed at, nothing is quietly dropped from the analysis, and everything else in the upload is still processed normally.
Can I send a folder of PDFs?
Yes, and it is the intended way to do a year of monthly reports. The structure worked out on the first is reused across the rest.