Opportunities › AI dictation and speech-to-text for knowledge workers › Proposed app
Hushnote
Private voice dictation that runs on your device, even offline
- Score
- 51
- confidence 39
- Evidence
- 4
- items on the page
- Rivals
- 1
- named in sources
- First version
- 35 d
- estimate
- Type
- Mobile app
- PWA: installs on iPhone and Android
- Market since
- 1997
- Mature market · AI estimate
ScreensConcept
Mobile app: Users hold to talk while on the go, in meetings, or away from a desk to capture notes and transcribe quickly. Sample screens on a phone, in the proposed brand. Their numbers are sample data.
Dictate
142 wpm
- ModelSmall, on device
- NetworkOffline
- LanguageEnglish
- Words today1,860
About this appAI proposal
Hushnote turns speech into text inside your browser. A speech recognition model is downloaded once and then runs on your own device, so dictation works without a connection and your audio is never uploaded.
It is made for people who write and take notes all day: students, researchers, consultants, managers. Hold a key, talk, and clean text lands in a note you own. Export to Markdown or copy into any other app.
Unlike cloud dictation, there is no account needed to start and no audio leaves the machine. Unlike closed tools, your notes are plain files you can take elsewhere. Paid plans add team vocabulary, sync and self-hosting.
- Tagline
- Say it once. Keep it on your device.
- Voice
- CalmPlainTrustworthy
- Why this name
- Hush signals quiet and privacy, since nothing is sent away; note names the plain output. Two simple words, easy to say.
In the first version
Members say which features they would use. It tells PWAS what people want; it changes no score.
Offline by default
The speech model is stored on your device after one download. Dictate on a plane or a poor connection, with no audio sent anywhere.
No member yet
Hold-to-talk capture
Press and hold a key or button to speak, release to insert text. Capture an idea in seconds without leaving the page.
No member yet
Custom vocabulary
Add names, jargon and acronyms so they are spelled right every time. Share lists with a team on paid plans.
No member yet
Clean-up pass
Optional local rules add punctuation, remove filler words and fix casing, so notes need less editing.
No member yet
Brand identityAI proposal
Check the name against trademark registers before using it.
Palette
#1F5F5BDeep Teal · primary#F6F1E7Paper · surface#1F2A2BInk · text#E07A3FEmber · accent#8FB39ASage · data
Typography
Aa
Hushnote
ABCDEFGHIJKLMNOPQRSTUVWXYZ 0123456789
Say it once. Keep it on your device.
Fraunces for headlines and numbers, Figtree for everything else.
PricingAI proposal
The material gives no competitor prices. These tiers are a starting point to test, based on the monetization inference of premium features, sync and enterprise licensing.
Free
Free
- Offline dictation on one device
- Small model
- Unlimited local notes
- Markdown export
Pro
$8 a month
- Larger, more accurate models
- Custom vocabulary
- Multi-language
- Meeting transcripts
- Encrypted sync
Team
$12 a month
- Per seat
- Shared vocabulary lists
- Admin controls
- Self-hosted sync option
What it could be worthEstimate
What a software business making this much sells for today, from real deals. Pick a monthly revenue.
If it makes
Worth about
$194,000
Sale range $132,000 to $263,000
$120,000 a year in revenue
From 87 software businesses listed and sold on Empire Flippers in the last year, revenue vetted by the broker: they sell for about 19× their monthly revenue, most between 13× and 26×.
Measured on profit, the middle deal is 32× monthly net profit: about 2.7 years of profit.
An estimate, not an offer or financial advice. A real price depends on profit, growth, churn, how much of the work is the owner's, and the buyer.
How it would growAI proposal
How the product would reach new customers through its own use, and how the first ones would be found. A plan to test, not a result.
- Built-in loop
- Shared outputs. Users export or share transcripts and notes as Markdown files or read-only links. A small optional 'Written with Hushnote' footer on shared pages shows where the text came from.
- Do they become customers?
- Recipients see the product but may not need dictation themselves. Test whether shared-link viewers click through and install, versus relying on search and content for sign-ups.
- First customers
- Graduate students and independent researchers who transcribe interviews and lectures and cannot upload recordings to cloud services.
- Where to reach them
- GitHub and open-source forums: Reach users of local Whisper projects with a no-install browser demo and a comparison post on privacy.
- University communities: Offer free Pro to student groups and research labs in return for feedback and referrals.
- Productivity communities: Share short offline-dictation demos in note-taking and writing forums.
- First milestone
- Twenty-five paying Pro users among researchers and students within 60 days of launch, with most using it offline.
Why people already pay for this
- Public claimDictation can be 3x faster than typing for capturing voice input.
- VerifiedGitHub repository gained 1552 stars in 520 days
- Public claimUsers want dictation that works offline and locally without sending data to cloud services.
- AI inferenceThere is active open-source community demand for this capability, as evidenced by repository engagement.
Tech stack
Every proposal builds on the same PWA base, installable from the browser on any phone or computer. Dashed boxes are what this app adds.
- AppNext.js and TypeScript
- PWAManifest and service worker
- UITailwind CSS
- Data and authSupabase Postgres
- PaymentsStripe
- EmailResend
- HostingVercel
- This app · On-device speech modelOpen-source Whisper-class model via WebGPU/WASMRuns recognition in the browser so audio stays local and works offline.
- This app · Local storageIndexedDB with origin private file systemStores model files and notes on the device for offline use.
RisksAI inference
- TechnicalBrowser inference of larger models can be slow on older laptops and phones, and model downloads are large. Accuracy comparable to cloud services must be tested on real devices.
- PlatformMobile browsers limit background audio, storage and WebGPU support, and iOS Safari may evict stored model data. System-wide dictation into other apps is not possible from a PWA.
- CommercialThe project is open source and free alternatives exist, so people may not pay. Willingness to pay for sync, vocabulary or team features is unproven.
- Intellectual propertyModel licences and any bundled vocabulary or language data must be checked before commercial use. The name needs a trademark check.
- RegulatoryRecording meetings may require consent of other participants in some places. Users need clear prompts and guidance.
Everything marked AI proposal or AI inference was written by a model from the evidence on the opportunity page. The proposed name must be checked against trademark registers before it is used.