Deposition transcripts as citable Markdown

What DepoScript is

DepoScript is a converter for deposition and hearing transcripts. It reads the transcript PDF and returns Markdown in which every line keeps the page and line number it was printed on, each question and answer is a separate turn, and any passage can be copied with a cite such as Smith Dep. 45:12–46:3.

It is made for the person who has to find testimony, quote it and cite it: a litigation associate, a paralegal, a lawyer working from the PDF the reporter sent. The promise is narrow. A transcript becomes citable Markdown, and every quote still knows its page and line. Checking the words against the certified transcript stays your job. The tool shows where the numbering looks wrong, so you know where to look first.

How it works

1. Upload the transcript PDF

   Drop the file the reporter sent into the workbench above, or press one of the two samples to see a result first. Full-size pages and condensed pages (several transcript pages on each sheet) are both read, and the Checks tab says which layout was found.

2. Read the checks before you quote

   Checks lists missing lines, short pages, repeated pages and page numbers that had to be inferred. A clean result means the numbering is consistent in the converted text. It does not prove that every word matches the PDF.

3. Choose how you want the text

   Numbered lines puts 2:11 in front of every line. Clean paragraphs joins each turn into one paragraph that starts with its range, such as [2:9–10]. Switching formats re-formats the text in your browser and does not use up another conversion.

4. Pick a passage and copy the cite

   Click a line number in the page view and then another, or type 45:12 and 46:3. The cite builder gives the cite, the quotation with its cite, and an inline quotation. Check each against the certified transcript before it goes into a filing.

What you get

The examples below were cut from the output of this site's own code, run on the full-size sample in the workbench. They are not typed by hand.

2:9 Q. Good morning, Mr. Smith. Please state your full
2:10 name for the record.

2:11 A. John Allen Smith.

2:12 Q. And where are you employed?

2:13 A. I am the director of operations at Northgate
2:14 Freight, LLC.

2:15 Q. How long have you held that position?

2:16 A. Since March of 2019. Before that I was a regional

[2:9–10] Q. Good morning, Mr. Smith. Please state your full name for the record.

[2:11] A. John Allen Smith.

[2:12] Q. And where are you employed?

[2:13–14] A. I am the director of operations at Northgate Freight, LLC.

[2:15] Q. How long have you held that position?

[2:16–17] A. Since March of 2019. Before that I was a regional dispatch manager for about six years.

A cite that follows the testimony across a page break

Q. Did you sign that agreement on behalf of Northgate?
A. I did, on June 3, 2022.
MR. BAKER: Objection. Form. You can answer.
THE WITNESS: I signed it after our legal department had reviewed the schedules.

(Smith Dep. 2:22–3:1)

The cite is Smith Dep. 2:22–3:1. The passage starts on page 2 and ends on page 3, and the range says so. The inline form is also offered: “Q. Did you sign that agreement on behalf of Northgate? A. I did, on June 3, 2022. MR. BAKER: Objection. Form. You can answer. THE WITNESS: I signed it after our legal department had reviewed the schedules.” (Smith Dep. 2:22–3:1)

Turns as JSON

[
  {
    "id": 3,
    "page": 2,
    "startLine": 9,
    "endPage": 2,
    "endLine": 10,
    "speaker": "Q",
    "kind": "q",
    "text": "Good morning, Mr. Smith. Please state your full name for the record.",
    "cite": "Smith Dep. 2:9–10",
    "examiner": "MS. OKAFOR"
  },
  {
    "id": 4,
    "page": 2,
    "startLine": 11,
    "endPage": 2,
    "endLine": 11,
    "speaker": "A",
    "kind": "a",
    "text": "John Allen Smith.",
    "cite": "Smith Dep. 2:11",
    "examiner": "MS. OKAFOR"
  }
]

Every turn has its first and last page and line, a speaker (Q, A or a named speaker), and its cite. Colloquy such as an objection is a turn of its own, so testimony and talk between counsel stay apart.

What it checks

Before you quote anything, the Checks tab reports on the numbering of the converted text:

• Whether pages run in order, and which pages are missing, as ranges.
• Whether each page runs from line 1 to the last line, and which pages are short or long.
• Whether a page appears twice. An identical copy is dropped; a copy that differs is kept and marked.
• Whether a page number was read from the page or inferred from its neighbours. Inferred numbers are marked.
• Which running headers and footers were removed, and which pages were set aside from the turns index as errata or certificate.

This is what it said about the condensed sample, 5 pages (61–65), 25 lines per page: 2 items to review:

• Page 64 appears 2 times with identical text. The extra copy was dropped.
• Page 63 has 24 of 25 lines (line 17 not extracted).

63:15 carriers and the paperwork had to be
63:16 tracked down.
63:17 [line not extracted]
63:18 you get to a stopping point.

Page 63 lost line 17 in the sample. The placeholder keeps lines 18 and after on their own numbers instead of letting every later line slide up by one.

Compared with the other routes

What it does not do

Plans

• Free: 1 file per day, up to 10 pages, no account.
• Pro, $79 a month: 60 transcripts a month, files up to 150 pages.
• Firm, $199 a month: 5 seats, 250 transcripts a month, files up to 150 pages.

No account on any plan: after payment a page code comes by email, you enter it once, and it stays in that browser. A 300-page transcript is over the current 150-page file limit. Split the PDF into parts of 150 pages or fewer and convert each part. Page numbers are read from the printed pages, so cites keep the transcript's own numbering; if a part prints no page numbers, set the first page number under Options. Pricing details.

Questions