The drug names
come out right.
Dictate "metoprolol succinate" into a general dictation tool and read back "metro pull all." The Medical edition ships Blabb with the medical dictionary already in it — built so drug names, clinical terms, and diagnoses land right the first time, with no training period and nothing sent to a cloud. A dictation aid, not a medical device — always review output before it enters the record.
Tens of thousands of terms, from sources you already trust
Named sources, not scraped lists
The dictionary is built from the FDA's own drug registry, clinical terminology from the National Library of Medicine, and ICD-10 diagnosis names. That's the vocabulary of the chart, the referral letter, and the discharge summary — spelled the way the profession spells it.
Catches what sounds right but isn't
The hardest errors aren't misspellings — they're real words that sound like the term you said: "loo poos" for lupus, "war far in" for warfarin. The spelling pass groups what was heard with dictionary terms that sound alike, then picks one only if it is close enough in spelling to be the same word. It works word by word, not on the meaning of the sentence, so it fixes near-misses rather than rewriting plain speech into jargon.
Two passes, two jobs
The edition works in two places, and they are not the same thing. The full 48,133-term dictionary drives the spelling pass over the raw transcription, fixing terms the speech model misheard. And a brief medical note goes into the cleanup prompt, so the AI editor knows it is handling clinical writing. It works with all six writing styles and alongside the personal glossary that learns your own corrections.
Spoken like a human. Typed like a chart.
"uh patient's on metro pull all succinate twenty five milligrams once daily for the a fib and um we'll check a TSH because of the fatigue"
Patient is on metoprolol succinate 25 mg once daily for AFib. Will check a TSH given the fatigue.
"started war far in five milligrams daily uh INR goal two to three and recheck in one week"
Started warfarin 5 mg daily. INR goal 2-3; recheck in 1 week.
Invented examples, real behavior: the drug names come from the dictionary, the numbers become digits, and the note stays exactly as long as you dictated it.
A dictionary is data. Data stays on your PC.
Your dictations never train a model
Most cloud dictation tools’ terms let them keep your session audio and notes, and train on them. Speech recognition and cleanup both run on your PC, and your audio and transcripts are never uploaded, so nothing you dictate can train any model — not Blabb’s, not anyone’s. You get the documentation help — without handing over what you’ve built.
Patient details never leave the machine
The Medical edition is exactly as private as base Blabb, because the dictionary isn't a cloud service — it's a term database installed with the app. Speech recognition, cleanup, and the dictionary lookup all run on your PC. Blabb reaches the network only to check your licence — once verified, it keeps dictating offline for 30 days before it needs to check again. The full picture: dictation for doctors.
History under your control
Dictation history is encrypted with keys tied to your Windows account, and retention is yours to set — don't save history at all, or 1 day, 1 week (the default), 1 month, or forever. One click deletes your history, saved words and logs.
Works where hospital IT locks down
Citrix, RDP, and VDI sessions with clipboard redirection disabled are normal in clinical environments. Blabb types key-by-key into the focused window, so dictation works inside them too. The one exception is a window running as administrator: Windows blocks synthetic input there, for every tool. How that works.
USD $9.99 a month — two coffees and change.
Everything in Blabb, plus the medical dictionary. Prefer once a year? USD $99.99 — two months free. Practice law instead? There's a Legal edition too.
FREE TIER 2,000 WORDS/DAY · 14-DAY UNLIMITED TRIAL IN-APP · CANCEL ANY TIME