Voice to Text Custom Vocabulary Mac: Adding Names, Jargon, and Acronyms

Wooden letter tiles spelling 'Pseudonym' on a table with a blurred background.

Photo by Markus Winkler on Unsplash

If you work in medicine, law, finance, or any field with specialized terminology, you already know the problem. Generic voice to text custom vocabulary on Mac software stumbles on drug names, case citations, ticker symbols, and client names it's never seen before. You spend more time correcting transcripts than you saved by dictating. This guide is specifically about fixing that — using VoicePrivate's custom vocabulary system to teach your local AI engine exactly the words it needs to know, so the output is useful the first time.

TL;DR

  • VoicePrivate lets you add custom terms directly in the app — no cloud sync, no account required, no data leaves your device.
  • Specialty editions (Healthcare, Legal, Finance, Insurance) ship with domain-specific vocabulary built in, giving you a head start before you add a single custom term.
  • Custom vocabulary in VoicePrivate works for both file transcription and live real-time dictation into other Mac apps.
  • All processing stays on-device, which means your custom word lists — including patient names, client names, or proprietary terms — never touch a server.

1. Understand What Custom Vocabulary Actually Does (and What It Doesn't)

Custom vocabulary tells the on-device speech recognition engine to expect specific words and to spell them a specific way. It doesn't retrain the underlying model. Think of it as a bias list: when the engine hears something that sounds like "metformin" or "Daubert" or "EBITDA", the custom entry pushes the probability toward the correct spelling instead of a phonetically similar common word.

In practice, this matters most for:

What custom vocabulary won't fix: background noise, a distant microphone, or heavily accented speech on a low-quality recording. Those are audio quality problems, not vocabulary problems. Accuracy varies by use case — vocabulary is one of the most controllable factors you have.

Note: VoicePrivate's custom vocabulary feature is available across all plans, including the free tier for basic transcription. Paid plans unlock additional features like speaker diarization, longer files, and specialty editions with pre-built domain dictionaries.

2. Choose the Right Edition Before You Start

If your work is in a specific domain, start with the matching specialty edition rather than General — it ships with domain vocabulary already loaded.

VoicePrivate has 5 editions:

Here's the thing: every term you don't have to add manually is time saved and one fewer potential typo in your custom list. A physician starting with the Healthcare edition is already working from a foundation that includes thousands of clinical terms. They only need to add what's specific to their practice — their hospital's department names, colleagues' names, the proprietary drug protocols their institution uses.

Specialty editions are available on paid subscription plans. You can review the full breakdown on the VoicePrivate pricing page.

Scrabble game spelling 'CHATGPT' with wooden tiles on textured background.

Photo by Markus Winkler on Unsplash

3. Add Individual Vocabulary Terms in VoicePrivate

Open VoicePrivate, go to Preferences, and find the Custom Vocabulary section — this is where you add terms one at a time.

For each entry, you provide two things:

  1. The display form - exactly how you want the word to appear in the transcript (e.g., "HbA1c", "Lipitor", "voir dire")
  2. The spoken form - how you actually say it out loud (e.g., "H-B-A-1-C", "lip-ih-tor", "vwar deer")

The spoken form is the piece that most guides skip over. If you only enter the display form, the engine has to guess how the word sounds. For medical abbreviations and Latin legal terms, that guess is often wrong. Write the spoken form phonetically — the way you'd explain the pronunciation to someone reading it cold.

Tip: For acronyms you speak as individual letters ("HIPAA", "LLC"), enter the spoken form with hyphens between letters: "H-I-P-A-A", "L-L-C". For acronyms you pronounce as words ("FICA" as "FYE-kah"), write it phonetically: "FYE-kah".

How Do I Add Words to Dictation on My Mac?

MacOS has two built-in options: Dictation (in System Settings > Keyboard > Dictation) and Voice Control (in System Settings > Accessibility > Voice Control). Both let you add custom vocabulary terms. For Dictation, the custom vocabulary option is limited. For Voice Control, you get more control: open Voice Control settings, click Vocabulary, and add terms with a display form and spoken form.

But neither Apple Dictation nor Voice Control was built for professional domain use. They work for occasional dictation. VoicePrivate was built specifically for power users who need consistent accuracy on specialized terminology, live dictation that types directly into any Mac app, and a local AI engine that processes everything without sending audio to Apple's servers.

4. Import Multiple Custom Vocabulary Terms at Once

If you have more than a handful of terms, don't add them one by one — use VoicePrivate's import function to load a prepared list.

VoicePrivate accepts vocabulary imports as plain text files. One term per line, with the display form and spoken form separated by a tab or delimiter. You can prepare this in any text editor or spreadsheet app, then import it directly.

This is the fastest way to get started with a large professional vocabulary. For example:

1
Prepare your list

Create a plain text file with one term per line. Include the display form and spoken form for each entry, separated by a tab character.

2
Open VoicePrivate Preferences

Navigate to the Custom Vocabulary section in Preferences.

3
Import the file

Use the import option to load your prepared list. All terms are added locally — nothing is uploaded.

4
Test with a short dictation

Dictate 2-3 sentences that use your new terms. Confirm the output matches what you expect before doing a long session.

Starter Vocabulary Lists by Profession

Below are starter term categories to include in your import file. Not exhaustive — a structured starting point you can extend.

Legal (priority terms to add):

Healthcare (priority terms to add):

Finance (priority terms to add):

Note: VoicePrivate's Healthcare edition is specifically designed for clinical environments. Learn more at the VoicePrivate Healthcare page. Because all processing is on-device, VoicePrivate is designed for HIPAA environments without a BAA — there is no data transmitted to protect on our end.

5. Modify and Delete Custom Vocabulary Terms

Incorrect entries are worse than missing entries — a wrong spoken form actively misdirects the engine.

In VoicePrivate's Custom Vocabulary panel, every term you've added is editable. Select a term, edit either the display form or spoken form, and save. If a term is causing consistent misrecognitions — meaning the engine is matching to it when it shouldn't — delete it and re-enter it with a more specific spoken form.

Common reasons to modify an existing entry:

Export your vocabulary list before making bulk changes. VoicePrivate lets you export the full list, which also makes it easy to back up your vocabulary or move it to a second Mac.

A decorative chalkboard displaying names at a refined indoor gathering.

Photo by Heiner on Unsplash

6. Use Custom Vocabulary with Live Dictation into Other Mac Apps

VoicePrivate's live dictation mode types directly into any Mac app in real time, and your custom vocabulary applies there too — not just in file transcription.

This matters more than most guides tell you. When you're dictating a note into your EHR, drafting a contract in Word, or writing a client email in Mail, the same custom vocabulary that powers your file transcription is active. You say "Daubert standard" and it types "Daubert standard" — correctly capitalized, correctly spelled. Every time.

Can I Dictate into Word on a Mac?

Yes. With VoicePrivate's live dictation active, you speak directly into Microsoft Word (or any other Mac app) and your words appear in real time. This works system-wide on macOS 13 and later. The custom vocabulary you've set up in VoicePrivate applies to every app you dictate into — Word, Pages, Slack, your email client, your EHR, any text field on your device.

This is different from Apple's built-in Dictation, which also works system-wide but has more limited custom vocabulary support and, depending on your settings, may send audio to Apple's servers for processing.

7. Per-App Transcription Modes and Vocabulary

VoicePrivate supports per-app transcription modes, which means you can configure different behavior for different applications.

In practice, this lets you set a more formal transcription style in your legal brief software versus a more conversational style in your messaging app. Combined with custom vocabulary, you can tune the experience app by app rather than using one-size-fits-all settings.

For example:

This level of control matters because vocabulary that helps in one context creates noise in another. "Discharge" means something very specific in a clinical note. Something different entirely in a legal letter.

8. Troubleshoot Custom Vocabulary That Isn't Being Recognized

If a term you added isn't appearing correctly in transcripts, the spoken form entry is almost always the cause.

Here's a systematic way to diagnose it:

  1. Check the spoken form, not the display form. Say the term out loud slowly. Write down exactly what you hear phonetically. Compare that to what you entered. A single phoneme difference can cause a miss.

  2. Look for conflicts with common words. If "axle" is being transcribed when you say a client's name "Axelle", the spoken forms are too similar. Add more phonetic specificity to your custom entry.

  3. Check audio quality first. Custom vocabulary can't compensate for a recording where the term is inaudible, clipped, or buried in background noise. Run a short test recording in a quiet environment to isolate whether it's a vocabulary issue or an audio issue.

  4. Re-export and review your full list. Sometimes an import adds duplicate entries or malformed rows. Export, open in a text editor, and check for obvious errors.

  5. Test with a dedicated file. Record a short audio file where you say only the problematic terms, then run it through VoicePrivate's file transcription. This isolates the term from any live dictation complexity.

Warning: Avoid adding extremely short spoken forms (one syllable) for multi-word display terms. The engine may match these too aggressively, inserting the term where it doesn't belong.

9. Privacy Considerations for Custom Vocabulary on Mac

Your custom vocabulary list contains sensitive information by definition — client names, drug names, proprietary terminology — and where that list is stored matters.

Here's how VoicePrivate handles it: your custom vocabulary lives entirely on your device. Never uploaded to a server. There's no account sync because there's no account. VoicePrivate requires no sign-in and collects no telemetry. The on-device speech recognition engine processes your audio and applies your vocabulary list locally, and none of it leaves the machine.

Cloud-based transcription tools are a different story. When you add custom vocabulary to a cloud service, both your audio and your vocabulary list travel to external servers. For professionals handling protected health information, privileged legal communications, or material non-public financial information, that's a compliance exposure — not just a privacy preference.

VoicePrivate is designed for HIPAA environments without requiring a Business Associate Agreement because there's nothing to protect on our end. The data never reaches us. For more detail on how this works, see our privacy architecture page.

Feature Apple Dictation Apple Voice Control VoicePrivate
Custom vocabulary Limited Yes Yes
On-device processing Optional (server mode exists) Yes Always
Live dictation into any app Yes Yes Yes
Domain-specific editions No No 5 editions
File transcription No No Yes (drag-and-drop)
Export formats None None .txt, .json, .md, .srt, .vtt
Speaker diarization No No Yes (paid plans)
No account required Yes Yes Yes

10. What Is the Best Voice to Text Software for Mac?

The best voice to text software for Mac depends on your use case — but for professionals who handle sensitive content and need domain accuracy, on-device processing with custom vocabulary support is the non-negotiable starting point.

Apple's built-in Dictation is convenient for casual use. Voice Control adds hands-free Mac control. Dragon for Mac has historically been the enterprise choice for dictation accuracy, though it carries significant cost and complexity.

VoicePrivate's position is specific: 100% on-device processing, 5 specialty editions with domain vocabulary built in, custom vocabulary that applies to both file transcription and live real-time dictation, and no account or internet connection required after the one-time model download on first run. Runs on macOS 13 and later, optimized for Apple Silicon, and also runs on Intel Macs.

For a broader look at how all of these capabilities compare, see Voice to Text Mac: Features, Speed, and Accuracy for Power Users.

Can Mac Convert Speech to Text?

Yes, macOS has two built-in speech-to-text tools: Dictation and Voice Control. Dictation is activated via System Settings > Keyboard > Dictation and converts spoken words to text anywhere you can type. Voice Control (System Settings > Accessibility > Voice Control) adds full Mac control by voice, including custom vocabulary support. Both have limitations for professional use — particularly around domain vocabulary, offline reliability, and export options. Third-party tools like VoicePrivate fill those gaps.

11. Offline Capability: Why It Matters for Custom Vocabulary

Once VoicePrivate downloads its local AI engine during first setup, it works completely offline — your custom vocabulary included.

Cloud-based voice to text services require an internet connection to function. On a plane, in a rural clinic, in a deposition room with restricted Wi-Fi — your transcription stops. With VoicePrivate, the model and your vocabulary list are both local. You dictate the same way whether you have full 5G or no signal at all.

And there's no latency from cloud round-trips. Your audio doesn't travel to a data center and back. The local AI engine processes it on your device's hardware, with Apple Silicon delivering fast transcription on M-series chips.

Can I Use Voice Dictation Offline on My Mac?

Yes, but it depends on the tool. Apple's Dictation has an "offline" toggle in System Settings (under Dictation), but it uses a more limited offline model. Apple Voice Control works offline by default. VoicePrivate is offline by design — the model is downloaded once and runs locally forever. No server mode, no fallback to cloud processing, no feature degradation when you're not connected. See the VoicePrivate features page for the full capability list.

Tip: If you work across two Macs (a desktop and a laptop, for example), export your custom vocabulary list from VoicePrivate on one machine and import it on the other. Since there's no cloud sync, this manual export/import is the way to keep both machines aligned.

Putting It Together: Your Custom Vocabulary Workflow

The practical sequence for getting the most out of voice to text custom vocabulary on Mac in VoicePrivate:

  1. Start with the specialty edition that matches your field (Healthcare, Legal, Finance, or Insurance) to get domain vocabulary built in.
  2. Prepare a vocabulary import file with the terms most specific to your practice — client names, proprietary terms, unusual proper nouns.
  3. For each term, write the spoken form phonetically rather than leaving it blank.
  4. Import the list, then test with a short dictation or file upload using sentences that include your new terms.
  5. Fix any entries that misfire by refining the spoken form.
  6. Export a backup of your final vocabulary list and store it locally.

Bottom line: 30 minutes of setup pays off in every session after that. The time you spend writing clean spoken forms for 50 specialized terms is returned to you the first time you transcribe a long document without correcting a single drug name or case citation.

For more on how VoicePrivate handles accuracy, speed, and professional workflows, visit Voice to Text Mac: Features, Speed, and Accuracy for Power Users. Questions about which plan is right for your use case are covered on the VoicePrivate FAQ page.

Key Takeaways

  • Custom vocabulary in VoicePrivate requires both a display form (how the word appears) and a spoken form (how you say it) — the spoken form is what most users skip and most errors trace back to.
  • VoicePrivate's 5 specialty editions (General, Healthcare, Legal, Finance, Insurance) include domain vocabulary out of the box, reducing the number of custom terms you need to add manually.
  • All custom vocabulary is stored and processed entirely on your device — no cloud sync, no server uploads, no account required. This makes VoicePrivate designed for HIPAA environments without a BAA.
  • Custom vocabulary applies to both file transcription and live real-time dictation into any Mac app, including Word, your EHR, email clients, and any other text field on macOS 13 or later.
  • After initial setup, VoicePrivate works completely offline — the local AI engine and your vocabulary list are both on-device, so transcription works the same with or without an internet connection.