Documents by type
1,491,796 documents sorted by what they print about themselves.
How these groups were made
Every document here was sorted by what it prints about itself: a header block, a clerk's stamp, a form number, a caption. No language model was used and no summary was written, so a group is only ever as good as the words on the page. Each document page shows the exact text that produced its label.
Roughly one document in twelve prints nothing a rule can recognise, and is listed under Unclassified rather than guessed at. Read every group below as documents identified as that kind, never as all of them. For a complete search of the text itself, use full-text search.
Every type
| Type | Documents | Share | Identified by |
|---|---|---|---|
| 1,103,954 | 74.00% | an anchored From / To / Subject / Date header block | |
| Unclassified | 121,576 | 8.15% | text was read, but it matched no printed form |
| No text read | 110,740 | 7.42% | no word-shaped text was read off the page |
| Bank record | 52,736 | 3.54% | a statement field such as Statement Period, Primary Account or Beginning Balance |
| Court filing | 46,901 | 3.14% | a clerk's stamp, or a court's own name at the head of the page |
| Pen register record | 7,017 | 0.47% | a PRTT record type and connection fields |
| Travel record | 5,794 | 0.39% | a travel agency invoice or itinerary header |
| Calendar or task item | 5,419 | 0.36% | a mail client's exported calendar or to-do block |
| Invoice | 4,645 | 0.31% | an invoice or billing header together with an amount |
| Device extraction report | 3,902 | 0.26% | a device extraction report's evidence id and section navigation |
| Flight record | 3,659 | 0.25% | a passenger manifest or flight log header |
| Letter | 3,333 | 0.22% | a salutation, or a Re: line with a closing |
| Memorandum | 2,489 | 0.17% | the printed word MEMORANDUM above a header block |
| Bureau of Prisons custody record | 2,423 | 0.16% | Bureau of Prisons or Metropolitan Correctional Center letterhead and inmate fields |
| Bank client due diligence | 2,409 | 0.16% | know-your-customer or due diligence report headings |
| Chat messages | 2,229 | 0.15% | an exported thread's Platform and Timezone header |
| Production placeholder | 1,891 | 0.13% | a slipsheet standing in for a file produced natively |
| Machine-readable export | 1,715 | 0.11% | XML or a delimited header row |
| Testimony transcript | 1,314 | 0.09% | the apparatus of testimony: an examination heading, counsel identifying themselves, a reporter's certification, alternating question and answer |
| Blank page | 1,273 | 0.09% | the production's own statement that the page is blank |
| FBI form | 1,238 | 0.08% | a printed FD form number, or an Electronic Communication lead block |
| Regulatory filing | 948 | 0.06% | a form number in a header line |
| Video transcript | 745 | 0.05% | the corpus record's own source field |
| Carrier phone records | 731 | 0.05% | Call Detail Records, Subscriber Information or a carrier's proprietary header |
| Handwritten document | 675 | 0.05% | the corpus record's own source field |
| Audio or video transcript | 423 | 0.03% | the corpus record's own source field |
| Corporate or trust instrument | 297 | 0.02% | the instrument's own name on a line of its own |
| Tax document | 258 | 0.02% | a tax form number in a header line |
| Agreement | 213 | 0.01% | the agreement's own name on a line of its own |
| Police report | 203 | 0.01% | an incident, offense or probable cause caption |
| Deposition | 157 | 0.01% | a DEPOSITION OF caption, or an exhibit cover naming a deposition beside the reporting firm |
| Fax cover sheet | 138 | 0.01% | a facsimile cover caption |
| Subpoena | 122 | 0.01% | a subpoena caption on a line of its own |
| Grand jury material | 118 | 0.01% | a grand jury caption at the head of the page |
| Hearing transcript | 45 | 0.00% | a TRANSCRIPT OF ... PROCEEDINGS caption |
| Proffer agreement | 41 | 0.00% | the printed words PROFFER AGREEMENT |
| Spreadsheet | 16 | 0.00% | the corpus record's own source field |
| Warrant | 6 | 0.00% | a warrant caption on a line of its own |
| Privilege log | 3 | 0.00% | a privilege log caption |
Two rules leave a row above without a link. A type with more than 4,000 documents is given as a count only, because a list that long is not something anyone reads: use full-text search for those. So are Unclassified, No text read, Production placeholder, Machine-readable export and Blank page, which record what is missing from a page rather than a kind of document, so a listing of them would be a page of ids with nothing to read.