Indexes what you already hold.
Every file becomes a row with its format, its domain and its sensitivity tier, whether it arrived as a spreadsheet, a recording or a photograph.
Back to Home
Product · Data RoomBook a walkthrough
Product · Data Room
The files your teams already keep, whatever shape they arrived in.
Spreadsheets, recordings and photographed forms, indexed and routed to the Desk that should read them.
Field capture
A form filled in at the line, and everything the machine read from it.
The photograph stays exactly as it arrived, with its checksum. Beside it sits the extracted text, a legibility figure that is a measure of the text rather than a model's confidence, and the candidate fields. A person corrects what is wrong and decides. Only that decision turns a capture into a verified record.
Why this matters
The evidence is not only what your office filed. It is the form a technician filled in at the gate, photographed on a phone, read by the machine and verified by a person before it becomes data.
What goes in
68 file formats, and what each one becomes.
Not a document library. The spreadsheets, recordings and photographed forms your teams already keep.
- Spreadsheets and tables
- Every column profiled and typed, each table a dataset with its lineage and the rows behind a figure.
- .xlsx .xls .xlsm .csv .tsv .parquet .numbers
- Documents
- Text, sections, dates and monetary amounts, each carrying the page it was read from.
- .pdf .docx .txt .md .rtf .odt .pages .pptx .key
- Captions and subtitles
- A timestamped transcript already, read on the parsing path at no transcription cost.
- .srt .vtt
- Archives and mail
- Opened and decomposed into what they hold, each attachment inheriting the boundary of the file it arrived in.
- .zip .7z .tar .gz .eml .msg
- AudioOn request
- A transcript, metered by duration. Over the plan's cap it is refused and says so, never quietly truncated.
- .mp3 .wav .m4a .amr .aac .flac .ogg .opus .bwf .mp2 .aiff
- VideoOn request
- The audio track transcribed, with the broadcast containers demuxed. The frames are not read, and nothing claims otherwise.
- .mp4 .mov .avi .webm .m4v .3gp .mxf .ts .m2ts .mpg .vob
- Photographs and scansOn request
- Text from a form filled in the field or a page put through a scanner.
- .jpg .png .heic .tiff .webp .bmp .avif .gif .svg
A format we cannot process yet is still named, with the reason and what it needs. Nothing arrives as unknown, and nothing is silently dropped.
What it does
An index in front of the Desks, not another place to file things.
The Data Room decides which source belongs to which Desk, and shows its working.
Recommends the Desk that should read it.
A licence goes to Compliance, a ledger export to the CFO Desk. The recommendation is shown for a person to accept; it does not move the file on its own.
Marks what needs a human.
A file the classifier is unsure of is held for review rather than routed, and says which domain it could not choose between.
Not yet
- Applying permissions from a classificationIn development
- A sensitivity tier is recorded and shown. Tightening access from it is an org admin action that is still being built, so nothing is restricted behind your back.
- The completeness checklistIn development
- Telling you which documents a due diligence or an audit is missing. The score exists as a column and is deliberately empty rather than estimated.
- Capture on a phoneOn request
- The capture, extraction and review chain runs, and it is enabled per organisation. The field app itself has no public release, so ask us rather than looking for it in a store.
Security
GDPR and CCPA compliant, ISO 27001 aligned, SOC 2 Type II in progress. Encrypted at rest and in transit. Your data never trains a model.
Security pageNext step
Bring the messiest folder you have.
A spreadsheet nobody owns, a recording of a site visit, a photograph of a form. We will index them with you and show you where each one lands.