Best Tools for Transcriptionists in 2026: The Complete Kit for Every Step of the Job

Short answer: There's no single best tool for transcriptionists. A working kit has eight parts:

1. Hardware. Good headphones, with an optional USB foot pedal.
2. Playback software. Express Scribe, The FTW Transcriber or oTranscribe.
3. A text expander.
4. An optional AI first draft, where your client allows one.
5. A style-guide check before you submit.
6. Proofreading and QA.
7. Subtitle QC, if you caption.
8. Delivery and business tools.

This guide covers what to look for in each tool, how to set it up, the mistakes people make, and when not to use it.

Most "best transcription software" lists are written for people who buy transcripts. This one is for people who produce them: freelance transcribers, legal-adjacent proofreaders and scopists, subtitle QC specialists, and small agencies.

How we chose

We grouped tools by the job they help with, not by brand. For every tool we checked its own website for what it does, which platforms it runs on, and whether your files stay on your computer. We leave out prices because they change often; check each vendor's site. We favored tools that:

• Fit how transcriptionists work: keyboard-first, built around client rules and editable files.
• Can run locally when confidentiality matters.
• Have a free version or trial you can test on a real file before you pay.

Disclosure: VideoText makes one tool in this guide, Make it client-ready (Guideline Format). We put it only where it fits, the style-guide step, and we list its limits the same way we list everyone else's.

Quick comparison: transcriptionist tools by step

Which kit fits your role

Step 1: Hardware (what equipment you need to be a transcriptionist)

You need a reliable computer, a comfortable keyboard, and good headphones. A foot pedal is optional. A laptop is fine. What matters most is that your playback software runs well on it and you can work at it comfortably for hours.

Headphones: the tool that matters most

You can't type what you can't hear. What to look for:

• Wired, closed-back over-ear headphones for blocking room noise. Under-the-chin transcription headsets are another option: Philips sells lightweight ones made for dictation and transcription.
• Wired over Bluetooth. A cable avoids Bluetooth lag and dropouts, which make hotkey rewinds feel off.
• Comfort for long sessions: soft ear cushions and low weight.

Setup tip: do a test on a hard file before a deadline: crosstalk, a quiet speaker, or a phone call. If you're cranking the volume, fix the audio (Step 8) rather than your ears.

Protect your hearing. The World Health Organization recommends keeping device volume at no more than 60% of maximum, taking regular breaks, and using noise-cancelling headphones so you don't need to turn the volume up in noisy places.

Foot pedals: optional but loved

A USB foot pedal gives you play, pause and rewind with your feet, so both hands stay on the keys. You don't need one to start, since hotkeys do the same job.

Check compatibility before you buy. It's the most common mistake people make with pedals:

• Express Scribe. NCH's compatible-controllers page lists the AltoEdge USB pedal as not needing a software license. Pedals such as the VEC Infinity IN-USB-1/IN-USB-2 need the paid license. So if you're on free Express Scribe, check that table before you buy.
• The FTW Transcriber. Its pedal page lists the Infinity USB-1, USB-2 and USB-3, Alto Edge, Olympus RS series and Start-Stop pedals.
• f4transcript. Its manual covers pedal setup, including Infinity models.
• Philips foot controls are built for Philips's own SpeechExec software. If you don't use SpeechExec, confirm support in your player first.

Setup tips:

• If your player has an auto-backspace setting (rewind a little on pause), turn it on so you hear the last word again when you resume.
• Put a non-slip mat under the pedal.
• If the pedal "isn't working", the usual causes are that the player doesn't support that model or that pedal control isn't switched on in its settings. Check both before you assume the hardware is faulty.

Keyboard, chair and screen

Transcription means hours of repetitive typing. OSHA's computer workstation eTool is a free, practical checklist:

• Top of the monitor at or just below eye level.
• Shoulders relaxed and elbows close to your body.
• Wrists in line with your forearms.
• Feet flat on the floor.

A second monitor helps if you keep the client's style guide open while you type.

Step 2: Playback and typing software (transcription software for transcribers)

This is the core of the job: controlling audio without taking your hands off the keyboard.

What to look for:

• Hotkeys you can remap.
• Variable speed without chipmunk distortion.
• An auto-rewind when you pause.
• Timestamp insertion.
• Support for your client's file formats and your pedal.
• Local playback, if files are confidential.

Express Scribe: the long-standing default (and its alternatives)

Express Scribe from NCH Software is playback software for typists, on Windows and Mac. It offers:

• variable-speed playback;
• hotkeys for mouse-free control while you type in Word or another word processor;
• professional USB foot pedal support;
• video playback.

Is Express Scribe free? There is a free version. NCH says it supports common audio formats, including WAV, MP3, WMA and DCT, and that the Pro edition adds proprietary dictation formats. It works with NCH's FastFox text expander.

Use it when you want a desktop player that handles dictation files and works with a pedal. Skip it when you need automatic timestamps every few seconds, since The FTW Transcriber is built for that.

Express Scribe alternatives:

• The FTW Transcriber for timestamps.
• oTranscribe for free use in the browser.
• Transcribe (Wreally) for a browser player with pedal support.
• f4transcript for interview transcripts.
• InqScribe for local AI drafts.

We compare them in more depth in our Express Scribe alternative guide.

The FTW Transcriber: built for timestamps

The FTW Transcriber is free-to-try Windows software. Its standout is automatic timestamps in your chosen format, starting from any timecode. It also has:

• per-client timestamp and hotkey settings you can save;
• hotkeys for phrases like "overtalking" or "unclear";
• a bookmark feature;
• playback from remote servers, with no download first;
• support for Infinity, Olympus, Philips and Grundig pedals.

Use it when clients want a timestamp every 30 seconds or at every speaker change. Skip it on a Mac, since it's Windows-only.

oTranscribe: free and in the browser

oTranscribe is a free, open-source (MIT) web app that puts the player and the text editor in one window. Its site says your audio and transcript never leave your computer, and your work saves to browser storage every second.

Shortcuts worth learning: Esc plays or pauses, and Ctrl+J inserts the current timestamp. You can export to Markdown, plain text or Google Docs.

Limits: it runs on desktop computers only, and its site doesn't list foot pedal support, so plan on hotkeys. Common mistake: clearing your browser data before you export, which deletes the work saved in browser storage. Export a copy as soon as you finish.

Transcribe (Wreally): browser player with pedal support

Transcribe lets you:

• slow down and auto-loop audio;
• use a foot pedal and a built-in text expander;
• keep working offline, since its editor and player run without an internet connection.

It also offers automatic transcription of uploaded files. Check your client's rules before you use that part (see Confidentiality).

f4transcript: for interview research

f4transcript, from audiotranskription.de, has:

• text modules, so an abbreviation plus a space inserts a phrase;
• timestamps that drop in automatically when you press Enter, or manually with F8;
• USB pedal support.

Microsoft Word set-up (your transcription template)

Most clients want DOCX, so set up Word once and reuse it:

• Build a template per client (.dotx) with their header block, fonts, line spacing and speaker-label style. Start every job from it.
• Use Word Styles for speaker labels and body text, so a global change takes one edit.
• Add AutoCorrect entries for labels and tags: File > Options > Proofing > AutoCorrect Options, then type the shortcut in Replace and the full text in With (Microsoft's steps).
• Learn Find and Replace with wildcards for repeat fixes. Review each match; don't use Replace All on the first try.

Common mistake: keeping one "master" template for every client. Labels, tag formats and timestamp styles differ, so one template leads to mixed formatting.

Step 3: Text expanders and shortcuts

You type the same speaker labels, tags and terms hundreds of times per file. A text expander turns a short abbreviation into the full text.

• PhraseExpress runs on Windows and Mac and has an iOS app. It installs locally, and its site names law offices and medical transcription as common uses.
• TextExpander expands snippets anywhere you type and supports fill-in fields. A team can share one snippet library, which helps agencies keep labels consistent.
• Word AutoCorrect costs nothing and works inside Word only.
• Built-in expanders come with some players: FastFox (with Express Scribe), and the expanders in Transcribe (Wreally) and f4transcript.

A starter snippet set:

Tips:

• Use a prefix that never occurs in real words, such as ; or //. Otherwise "in" turns into "Interviewer:" mid-sentence.
• Keep one snippet group per client, because their tags differ ([inaudible] vs [unintelligible] vs (inaudible 00:04:12)).
• Sync with care. If a client forbids cloud storage of their terms, use a local-only expander.

Step 4: AI first draft, then human edit

Many transcriptionists now start from a machine draft and spend their time correcting it. Others type from scratch, and some platforms require it. Check your agreement first (see Confidentiality).

Options for a first draft:

• InqScribe says its AI transcription runs on your computer. It combines a player and an editor with reusable snippets and clickable timecodes, and it exports SRT, WebVTT and SCC.
• MacWhisper says its transcription runs on your Mac, so no data leaves your machine. It also offers optional cloud services, so check which mode you're using. It exports DOCX, SRT and VTT, among other formats.
• Transcribe in Word works in Word for the web with Microsoft 365. You upload WAV, MP4, M4A or MP3, and it separates speakers. There's a monthly limit on uploaded minutes.
• Otter records from a Mac or Windows desktop, a phone or a browser. It's built around meetings, so you do the client's formatting yourself.
• Descript lets you edit audio and video by editing text, and it can remove filler words. It's strongest for podcast and video producers.
• VideoText Video to Transcript produces an editable transcript with timestamps per interval, segment or speaker. It exports TXT, DOCX, PDF, SRT, VTT, JSON or CSV. Free accounts can upload files up to 30 minutes.

How to edit an AI draft:

1. Set the rules before you touch the text: verbatim level, label format, tags, timestamps. Editing first and formatting later means doing two passes over everything.
2. Build a term list of names, places, products and acronyms from the job notes, then search the draft for each one.
3. Fix speaker attribution at every turn change and overlap. Machine drafts often merge two speakers.
4. Listen to the hard parts, not just the text that looks wrong: crosstalk, accents, numbers, negations ("can" vs "can't") and quiet speakers.
5. Mark what you can't hear with the client's tag. Don't guess.
6. Apply the style guide (Step 5), then proofread (Step 6).

When not to use an AI draft:

• when the agreement forbids outside tools;
• when the audio is so poor that correcting the draft would mean retyping most of it;
• when the job is strict verbatim and the draft tool cleans up fillers by default.

More on this: why AI transcripts fail QA and how to tell when an AI transcript is clean enough to deliver.

Step 5: Follow the client's style guide

This is the step most tool lists skip, and it's where a lot of rejections happen. Every client or platform has its own transcription style guide, and getting the words right isn't enough if the format is wrong.

Turn every guide into a one-page rule sheet before your first file. These are the categories almost every guide covers:

Verbatim levels in one line each. For more depth, see clean verbatim vs full verbatim.

• Strict / full verbatim: everything, including "um", "uh", stutters, false starts and, often, non-speech sounds.
• Clean verbatim: the same words and meaning, without fillers, stutters and false starts that add nothing.
• Edited / intelligent verbatim: lightly corrected grammar for readability. Only use it when the guide explicitly asks for it.

Want a ready-made sheet? Our freelance transcription style guide cheatsheet maps typical style-guide rows to what to check. For speaker labels you need to fix later, see renaming speaker labels in an existing transcript, and for missing timecodes, adding timestamps to a transcript you already have.

Worked example: applying a style guide to a real transcript

Here's a short excerpt and a fictional client guide, applied rule by rule.

The client's rules (example):

1. R1. Clean verbatim: remove fillers (um, uh), stutters and false starts that don't change the meaning.
2. R2. Speaker labels: by role, in bold, followed by a colon: Interviewer: / Respondent:
3. R3. Timestamps: [hh:mm:ss] at the start of every speaker turn.
4. R4. Tags: lowercase, in square brackets. Inaudible sections carry the time: [inaudible 00:04:24]. Overlapping speech is [crosstalk].
5. R5. Numbers: spell out one through nine and use numerals from 10. Spell out any number that starts a sentence. Write percentages as a numeral plus "percent".
6. R6. Spelling: US English. Use the client glossary for names ("Acme Health").
7. R7. Layout: one paragraph per turn, a blank line between turns, sentence case and end punctuation.

Before (draft):

00:04:12 S1: so um how many sites did you, did you visit last year
00:04:15 S2: uh twelve i think. maybe 12. we went to- we went to three in ohio and the rest were (???) and the response rate was about forty five per cent
00:04:31 S1: and that was funded by acme health
00:04:33 S2: yeah [overlapping] yes it was

After (client-ready):

[00:04:12] **Interviewer:** So how many sites did you visit last year?

[00:04:15] **Respondent:** Twelve, I think. Maybe 12. We went to three in Ohio, and the rest were [inaudible 00:04:24]. And the response rate was about 45 percent.

[00:04:31] **Interviewer:** And that was funded by Acme Health?

[00:04:33] **Respondent:** Yeah. [crosstalk] Yes, it was.

What changed, and who has to check it:

That split is the point. About half of style-guide compliance is mechanical rule-following, and that's where people slip on a deadline. The other half needs a human listening. A good workflow lets a tool or checklist handle the first half and leaves your attention for the second.

Make it client-ready (Guideline Format) by VideoText

Make it client-ready applies a client's style guide to a transcript you've already finished, then shows you what it did.

1. Add the rules. Upload the client's style guide (PDF, DOCX, TXT or Markdown, up to 8 MB), and the tool lists the rules it finds. Or start from an editable preset based on common transcription-company style guides. You can edit any rule before you run it, and undo an edit.
2. Add the transcript. Upload TXT, SRT, VTT or DOCX up to 3 MB, or paste the text. It can come from any tool, including one you typed yourself in Express Scribe.
3. Review the per-rule report. Each rule is listed as applied, needs a look, or couldn't be applied. Changes the tool isn't sure about are flagged, and you can accept, reject or undo each one, or accept them all. For SRT and VTT files it also checks subtitle limits.
4. Compare, edit, export. Compare the original with the formatted version, edit anything you want, then export TXT, DOCX, PDF, RTF or JSON, or the flagged sections as CSV. If you started from SRT or VTT, you can export SRT or VTT too.

What it doesn't do:

• It formats only. It doesn't re-transcribe the audio or check that the words are right. In the worked example above, it can't hear whether 00:04:31 is a question, or find the time for an inaudible tag. Those are yours.
• It doesn't replace timing QC for subtitles (see Step 7).
• There's no plain-language box for typing your own rule. Rules come from the uploaded guide or a preset, and you edit them there.
• It needs an editable file. If you proofread court transcripts as annotated PDFs, it isn't for you.
• Check your agreement first. It's an online tool, so the confidentiality rules apply.

Plans: you need a free account to see the result. Free accounts get 1 run a day, on transcripts up to about 30 minutes of text. Pro removes the daily cap and the 30-minute cap.

Other ways to check style compliance

• A personal checklist built from each client's guide and run before every submission. It's free and it works, but it's easy to skip on a deadline.
• Word macros and wildcard Find and Replace for repeat fixes like tag formats. They're powerful, but you build and maintain them per client.
• PerfectIt (next step) for consistency against a house style in Word.

Step 6: Proofreading and transcript QA

Proofreading is a separate pass from typing. Do it after formatting, so you check what the client will actually receive.

Tools

• PerfectIt scans a whole document for inconsistent spelling, numerals, hyphenation and capitalization, and checks it against a style guide. Its Windows edition needs no internet connection, so documents stay on your computer. The Mac-compatible version is cloud-based. It checks the document, not the audio.
• Word's built-in tools: spelling and grammar, Track Changes for client revisions, and Read Aloud, which helps you hear the typos your eyes skip.
• Your player's bookmarks (The FTW Transcriber has them, and so do InqScribe's clickable timecodes) to jump back to spots you flagged.

A transcript proofreading checklist (in this order)

1. Completeness. The transcript starts and ends where the audio does. There are no jumps, missing sections or duplicated blocks.
2. Speakers. One consistent label per person. Check every turn change and every overlap.
3. Terms. Names, places, products and acronyms match the term list or glossary everywhere.
4. Numbers, dates, negations. Re-listen to every one that matters ("fifteen" vs "fifty", "can" vs "can't").
5. Tags. Every inaudible, crosstalk or phonetic tag is in the client's exact format, with a time where required.
6. Timestamps. Right format, required interval, in order, and spot-checked against the audio.
7. Verbatim level. It's the same throughout. Watch for clean-verbatim slips in a full-verbatim job, and the reverse.
8. Layout and file. The template, header block, file name and export format are all right.
9. Final listen to every section you flagged.

For the full process, see what transcript QA is and the 7-step QA review workflow.

Common transcript QA mistakes

1. Speaker drift. "Speaker 2" becomes "Speaker 3" halfway through a long file, or two voices get merged.
2. Guessing instead of tagging. A plausible word where the audio is unclear is worse than an honest [inaudible].
3. Mixed tag formats in one file: [inaudible], (inaudible) and [unintelligible] side by side.
4. Timestamps in the wrong format or interval, or drifting away from the audio after an edit.
5. Numbers that follow no rule, like "12" in one paragraph and "twelve" in the next.
6. The wrong verbatim level, for example removing "um" from a strict-verbatim legal file.
7. Over-editing. "Fixing" a speaker's grammar or word choice when the guide says to keep it.
8. Find-and-replace damage, such as Replace All turning "in" into "Interviewer:" inside words.
9. Proofreading before formatting, and then never re-reading the formatted version.
10. A wrong template or file name, like last client's header or an unfilled case name or job number.
11. Spell-check "corrections" that change proper nouns or technical terms.
12. Skipping the final listen on the sections you flagged yourself.

Mistakes 3–6 and 10 are mechanical, which is where a checklist or a formatter helps most. Mistakes 1, 2, 7 and 12 need your ear.

Step 7: Subtitle QC

Caption and subtitle QC is a different job from transcript formatting. You check timing (when each cue appears and disappears), reading speed, line length and line breaks, on top of the text.

Tools:

• Subtitle Edit is free and open source for Windows, Linux and macOS. It handles 380+ subtitle formats and shows a waveform and spectrogram for precise timing, and it can fix common errors and spell-check.
• Aegisub is a free, cross-platform, open-source subtitle editor focused on timing subtitles to audio and styling them.
• VideoText's browser checks: Fix Subtitles for overlaps, reading speed and line length in an SRT or VTT, plus the subtitle reading speed checker and the SRT validator.

Know the client's spec. Platforms publish their own rules. Netflix's Timed Text Style Guide: General Requirements, for example, sets a maximum of two lines, a minimum duration of five-sixths of a second and a maximum of 7 seconds per subtitle event. Language-specific guides cover characters per line. For educational captioning, the DCMP Captioning Key covers speaker identification, presentation rate and more. For the numbers by style guide, see our CPL and reading-speed guide.

Where Make it client-ready fits: it can apply a text style guide to an SRT or VTT file and check subtitle limits, but it doesn't replace timing QC. Do timing in a subtitle editor.

Common mistakes:

• Fixing text in a word processor and breaking the cue numbering or timecodes.
• Merging cues without re-checking reading speed.
• Delivering SRT when the client asked for VTT, or the reverse.

Step 8: Delivery, file conversion and business tools

Deliver exactly what the client asked for:

• DOCX for most general, legal-adjacent and research clients.
• TXT for platforms that paste text into their own systems.
• SRT or VTT for captions. The WebVTT format is a W3C specification.
• PDF for a read-only copy.

Before you send, check three things: the file name follows the client's convention, the speaker labels match the guide, and the template header (case name, date, job number) is filled in.

Helpful utilities:

• VLC is a free, open-source media player for files your transcription player won't open.
• HandBrake is an open-source video transcoder for converting odd video formats.
• Audacity is a free, open-source audio editor for Windows, macOS and Linux. Use it to boost a quiet track, reduce noise, convert formats or find a quiet word in the spectrogram view. Work on a copy, never on the client's original.
• Fix Transcript corrects an existing transcript's text, speakers and timestamps without retranscribing.

Business tools:

• Time tracking such as Toggl Track or Clockify, so you know your real hourly rate per client and file type.
• A simple job log (spreadsheet) recording each job's client, audio length, due date, guide version and invoice status.
• See also how to earn more per hour as a transcriptionist.

Legal transcription and scopist tools

Legal work has its own stack, and the client usually decides it for you.

• CAT software (computer-aided transcription) is the court reporter's platform. It translates steno into a rough draft that a scopist edits. Two of the main ones are Stenograph's Case CATalyst, which sells an Edit edition and has a scopists section, and Advantage Software's Eclipse, whose Connection Magic team editing lets multiple scopists and proofreaders work in a reporter's document over the internet. If you scope, use what your reporters use.
• Digital-recording transcription for legal clients usually runs on a player like Express Scribe or The FTW Transcriber, sometimes with a firm-mandated portal. The AAERT (American Association of Electronic Reporters and Transcribers) offers certification and education for digital reporters and transcribers.
• Helpers: a local text expander for legal phrases, a legal dictionary, a per-case term list, and PerfectIt for consistency.
• Where Make it client-ready fits: only when you receive an editable DOCX or TXT and your agreement allows uploads. It isn't for annotated-PDF proofreading.

For context on the profession, the U.S. Bureau of Labor Statistics' court reporters and simultaneous captioners profile covers duties, training and licensing. For depositions and legal interviews, see our legal transcription workflow guide.

Confidentiality: read this before you upload anything

Some transcription platforms ban third-party online tools outright. Speechpad's worker FAQ, for example, says none of the tools you use should require or allow media or text to be uploaded to a third-party website. It also says local copies must be erased after you submit. Legal and medical clients often have similar rules.

Before you use any online tool, including VideoText, read your agreement. When uploads aren't allowed, stay local:

• Playback: Express Scribe, The FTW Transcriber, f4transcript, or oTranscribe (which keeps your files on your computer).
• AI drafts on your own machine: InqScribe or MacWhisper.
• Proofreading: PerfectIt for Windows.
• Text expansion: a locally installed expander.

Also delete files on schedule, don't sync client audio to personal cloud folders, and keep each client's style guide private.

Your complete transcriptionist kit (checklist)

Copy this into your notes, or print it, and tick it off as you build your kit.

Hardware

• Computer that runs your playback software smoothly (a laptop is fine)
• Wired closed-back or under-chin transcription headphones
• USB foot pedal checked against your player's compatibility list (optional)
• Keyboard, chair and monitor set up per the OSHA workstation checklist
• Second monitor for the style guide (optional)

Software

• Playback: Express Scribe, The FTW Transcriber, oTranscribe, Transcribe (Wreally) or f4transcript
• Word processor with one template (.dotx) per client
• Text expander with one snippet group per client
• First-draft tool, only where the client allows it
• Style-guide check: a per-client rule sheet, Make it client-ready, or both
• Proofreading: PerfectIt and/or Word's spelling, grammar and Read Aloud
• Subtitle editor if you caption: Subtitle Edit or Aegisub
• Audio and file utilities: Audacity, VLC, HandBrake

Per client

• Style guide saved and turned into a one-page rule sheet
• Term list or glossary
• Label, tag and timestamp snippets
• Template with header block and file-name convention
• Confidentiality rules noted: uploads allowed? deletion deadline?

Before every submission

• Rules applied (verbatim, labels, timestamps, tags, numbers, spelling, layout)
• Proofreading checklist done (Step 6)
• Final listen to every flagged section
• Correct format, file name and header
• Local copies deleted if the agreement says so

Starting out: practice and building a business

• Practice on real audio with free tools first (oTranscribe or free Express Scribe). Learn your player's hotkeys before you buy a pedal.
• Typing practice: a general typing test such as Monkeytype builds keyboard fluency. Audio typing tests, where you type from a recording, are closer to the real job. Many platforms run their own transcription test with their own style guide, so read that guide before you take it.
• Learn the style-guide skills, not just the typing: verbatim levels, tags and speaker labels. They're what separate a passable file from a client-ready one.
• Pick a lane. General, research, legal-adjacent and subtitle work each need slightly different tools (see the role table).
• Starting a transcription business needs a few tools beyond the kit: time tracking, a job log, a professional email, invoicing, and written terms that spell out delivery times, file formats and confidentiality.
• Looking for transcription jobs? For each platform, check whether its agreement allows AI drafts or online tools before you build your workflow around them.

Glossary

• Clean verbatim: the speaker's words and meaning without fillers, stutters and false starts.
• Full / strict verbatim: everything as spoken, including fillers, stutters and, often, non-speech sounds.
• Speaker label: the name or role that starts each turn, for example Interviewer:
• Timestamp: a time marker such as [00:04:12], placed at an interval or at each turn.
• Tag: a bracketed note such as [inaudible 00:04:24], [crosstalk] or [laughter].
• Crosstalk: two or more people speaking at once.
• Style guide: a client's rules for verbatim level, labels, tags, numbers and layout.
• Foot pedal: a USB device for controlling playback with your feet.
• Text expander: software that turns a short abbreviation into longer text.
• Scopist: an editor who turns a court reporter's CAT draft into a finished transcript.
• CAT software: computer-aided transcription software used by stenographic court reporters.
• SRT / VTT: common subtitle file formats with numbered or timed cues.
• CPS / CPL: characters per second (reading speed) and characters per line in subtitles.
• Subtitle QC: checking subtitle timing, reading speed, line length and text against a spec.

More terms: the VideoText transcription and subtitle glossary.