OCR for Court Reporters: Exhibits and Older Transcripts

OCR for court reporters is about everything around the transcript. Your transcript is text from the first keystroke, but the exhibits marked at the deposition, signed errata sheets and the paper transcripts in your archive are images. OCR makes those searchable as working copies, while the certified transcript stays exactly as you certified it.

This post is written for the freelance reporter, the official reporter with a storage room, and the production staff at a reporting firm. It stays with the documents your business handles and leaves trial strategy to the lawyers.

Your certified transcript stays the record. Use conversion on the deposition exhibits, the errata, and the older paper volumes a lawyer asks you to search. A court reporter can hand the attorney a working copy and still point the judge or the witness back to the certified pages. Store those working files under your records retention rule, and keep them apart from the transcript you signed.

Where does a court reporting firm still meet scanned paper?

A court reporting firm still meets scanned paper in four places: exhibits handed across the table, read-and-sign paperwork that comes back by mail or fax, the back catalog of transcripts from before digital delivery, and the firm's own job paperwork. The transcript is the one document in the job that was never paper first.

Exhibits shared on screen

When exhibits are shared on screen, some arrive as files with real text and some as flat scans the lawyer made that morning. You cannot tell which from the file name. A set of twenty exhibits from one deposition can be half searchable and half not, which is the worst case for the attorney who assumes a search covered all of them.

The errata loop

The federal rule gives a deponent who asks for review 30 days after being notified that the transcript is available to review it and sign a statement listing changes and the reasons for them, and the officer must attach any changes made in that period. That is in Rule 30(e) of the Federal Rules of Civil Procedure; state practice differs. The statement comes back handwritten, often scanned or faxed. Recognition is weak on handwriting, so OCR helps here mainly with the typed parts of the sheet: finding which witness and which job it belongs to when a stack arrives without cover letters.

Do court reporters need OCR when transcripts start as text?

Court reporters do not need OCR for a transcript they produced themselves, because the steno software's output is already text. They need it for the documents that travel with the transcript and for transcripts that survive only on paper. The need sits in production and archive work, not at the machine.

When you never need it

If you still hold the job file from your reporting software, export from that. A PDF, a plain text file or a condensed version generated from the source is exact. Running OCR on a printout of a transcript you could re-export adds errors to a document that had none.

When you do

• The source file is gone: old media, a retired system, a reporter who left the firm.
• The transcript came from another reporter or firm and you hold only a paper or image copy.
• A client sends a scanned transcript and asks for a word index or a searchable copy.
• The document was never a transcript: exhibits, errata, correspondence.

A two-minute test for any PDF

Open the file and try to highlight one word in the middle of a page. Then search for a word you can see. If neither works, the page is an image. Check an early exhibit and a late one, because bundles are often assembled from several sources.

What a lawyer means by "searchable"

When an attorney asks whether the exhibits are searchable, the practical meaning is that typing a name into the PDF viewer finds every exhibit page where it appears. That is a text layer behind the image. It does not mean the exhibits have been indexed, summarized or linked to testimony, and it is worth saying so on the order form so nobody expects more than they bought.

Making marked deposition exhibits searchable before delivery to counsel

Making marked deposition exhibits searchable before delivery to counsel means scanning each exhibit after it is marked, converting it to a searchable PDF that keeps the page image and sticker intact, naming it by exhibit number, and spot-checking before it goes out with the transcript. The image is the exhibit. The text layer is a convenience.

Under the federal rule, documents produced for inspection during a deposition must, on a party's request, be marked for identification and attached to the deposition, as Rule 30(f)(2) on documents and tangible things puts it. How exhibits are kept and delivered in state cases depends on that state's rules and on what counsel stipulate.

A production procedure

1. Scan after marking, not before. The sticker and any handwriting the witness added on the record are part of the exhibit.
2. One file per exhibit. Do not merge the set into one PDF before conversion; you will only have to split it again.
3. Scan in color when the exhibit has color, such as highlighted passages, photographs or a plan with colored lines.
4. Convert to searchable PDF. Choose the export that keeps the original page image with text behind it, so nothing about the exhibit's appearance changes.
5. Compare page counts. A 46-page exhibit must come back as 46 pages.
6. Name by a fixed pattern, for example Ruiz depo 2026-09-30 - Ex 014 - service invoice.pdf. Pad the number so files sort correctly.
7. Spot-check each file. Search for one word from the first page and one from the last.
8. Deliver with an exhibit list that shows exhibit number, description and page count.

A worked example

At the fictional deposition of Elena Ruiz in Pinder v. Castellan Mechanical, counsel mark 23 exhibits. Fourteen were emailed beforehand as PDFs; nine are paper handed over in the room, including a 46-page service log with handwritten technician entries.

Back at the office, the production coordinator tests the fourteen. Eight have text. Six are flat scans. Those six and the nine paper exhibits, fifteen in all, go through conversion. The service log's printed column headings and typed dates come through; the technicians' handwriting mostly does not. The delivery note to counsel says so in one line: "Exhibits 1 to 23 delivered as searchable PDF. Handwritten entries in Exhibit 14 are not reliably searchable."

That one line matters. An attorney who searches the set for a technician's name and finds nothing needs to know the search could not see the handwriting.

Things that get in the way

Exhibit stickers placed over text hide what is under them from any software. Place stickers in a margin when you have the choice. The post on OCR for trial exhibits deals with stickers and labels in more detail.

Oversized pages, spreadsheets printed at tiny scale and photographs of documents all convert less well. Do not promise results on those; tell counsel which exhibits they are.

Uploads to Schema Law are PDF, PNG or JPEG files of up to 20 MB each, so a long color exhibit may need splitting, and the per-document page limit depends on the plan. The site's how it works section describes the steps.

Confidential designations

Exhibits in a case with a protective order may be marked confidential or for attorneys' eyes only. Before any exhibit goes to an outside conversion tool, check what the order and your engagement terms allow, and ask the noticing attorney when unsure. The post on whether it is safe to upload client documents to an OCR tool lists the questions to put to a vendor.

Converting archived paper transcripts from older cases

Converting archived paper transcripts from older cases is worth doing when someone is likely to ask for them again and the source files are gone. Scan the paper as it is, convert to a searchable PDF, and label the result as an uncertified working copy so it is never mistaken for the transcript you certified.

Decide what to convert

Not every box earns the effort. Start with your retention obligations, which depend on your state, your court if you are an official reporter, and your contracts. Then sort:

Scanning transcript paper

Transcript pages are friendly to recognition: one typeface, generous spacing, numbered lines. The trouble comes from the format around the text. Line numbers down the left margin can be read as part of the testimony. Condensed transcripts print four small pages on a sheet, and the reading order across those quadrants can scramble. Bound copies curve at the spine.

Remove the binding if your retention rules let you, feed pages flat, and scan the full-size version in preference to a condensed one when you hold both. The post on making deposition transcripts searchable explains what happens to page and line numbers after conversion and how to check them, so this post leaves that subject there.

Label the output

Add "uncertified working copy, converted from scan" to the file name and keep the scanned image inside the PDF. Keep the original certificate page in the scan. Anyone opening the file years later should be able to see what it is and where it came from.

For the box-by-box mechanics of a scanning project, the guide to digitizing law firm paper files applies to a reporter's storage room as much as to a firm's.

Does an OCR copy change the status of a certified transcript?

An OCR copy does not change the status of a certified transcript. The certified transcript is the one you certified, in the form you delivered it. A searchable copy made later from a scan is a working copy, and it carries no certification of its own. Rules on certification and on electronic transcripts vary by jurisdiction.

In federal depositions, Rule 30(f)(1) requires the officer to certify in writing that the witness was duly sworn and that the deposition accurately records the witness's testimony, and the certificate must accompany the record. Rule 80 allows stenographically reported testimony to be proved by a transcript certified by the person who reported it. Both tie the transcript's standing to the reporter's certificate. Software that guesses at characters in a scan is not part of that chain.

State statutes, court rules and licensing boards set their own requirements for certified copies, electronic signatures and who may produce a transcript. This post does not tell you what your jurisdiction requires. Check your board's rules and the court's.

Keeping the two apart in practice

• Never overwrite. The certified file and the scan of a certified paper original are stored read-only. Converted copies are separate files.
• Name the difference. Put "working copy" in the file name of anything produced by OCR.
• Do not attach your certificate to converted text. If a client needs a certified copy, produce it the way your rules provide, from the original.
• Tell the client what they have. One sentence on the delivery note is enough: the searchable copy is for locating testimony, and citations should be confirmed against the certified transcript.
• Do not correct the text layer by hand and pass it off as the transcript. A corrected working copy is still a working copy.

When a lawyer asks for a certified copy of an old transcript

If the source file exists, your normal process applies. If only paper exists, what you may certify and how is a question for your jurisdiction's rules, and possibly for the court. A searchable scan helps the lawyer find the testimony in the meantime. It does not answer the certification question.

Offering searchable exhibits as a court reporting service

Offering searchable exhibits as a court reporting service makes sense when your clients already ask for it, your production staff can add the conversion and checking steps without delaying transcripts, and you can describe the deliverable in one honest sentence. Decide whether it is included in exhibit handling or listed as its own line.

Work out the volume

Count exhibit pages for a typical month. A firm covering 60 depositions a month at an average of 150 exhibit pages handles about 9,000 pages. Then subtract what already has text. If a third arrives searchable, about 6,000 pages need conversion.

Match that to a plan. With Schema Law's published plans, 6,000 pages a month is beyond the Solo allowance of 1,000 and within the Team plan's 10,000, which also takes documents of up to 1,000 pages. Each account is for one person, so that assumes one production coordinator does the converting, and a second person would need a second account. A solo freelance reporter converting a few hundred exhibit pages a month is inside Solo. The pricing page has the details, and the post on OCR cost per page shows how to turn a plan into a per-page figure you can price against.

Describe the deliverable carefully

Write the service description so it promises what you control:

Exhibits delivered as PDF with a text layer for searching. Page images are unchanged. Handwriting, stamps and poor copies may not be searchable. The searchable text is not certified.

Do not promise accuracy. You did not create the exhibits and cannot control the quality of what counsel hands you.

Pricing and disclosure

How you charge is a business decision bounded by your contracts and by any rules in your state on reporter fees and on equal terms for all parties. Whatever you choose, show it on the rate sheet and the invoice so every ordering party sees the same thing. The post for records retrieval companies covers the same decision for a neighboring service business, in its section on charging law firm clients for OCR on retrieved records.

Protect the turnaround

The transcript deadline does not move because exhibits are being converted. Put conversion in the production checklist right after exhibit scanning, give it to a named person, and set a rule for expedited jobs: if conversion would hold up delivery, send the transcript and image exhibits on time and follow with the searchable set.

Start small

Offer it to three regular clients for a month. Track pages converted, staff minutes per job and any complaints. If the minutes are low and the clients keep asking, add it to the rate sheet.