What model does VoiceDash use?
VoiceDash cites an "OpenAI partnership" but does not name the speech model it runs. It publishes no accuracy benchmark, only a "98% user satisfaction rate," so we cannot compare it with Avalon.
| Feature | Aqua Voice | VoiceDash |
|---|---|---|
| Speech model | Avalon, Aqua's own proprietary model | Third-party API (cites an OpenAI partnership; specific model not published) |
| Public benchmark | 97.4% on AISpeak-10; #1 proprietary at OpenASR debut (Oct 2025) | None published (cites '98% user satisfaction') |
| Technical-term accuracy | Strong; adapts to on-screen names, code, and jargon | Adapts formatting to the app you're in |
| Real-time output | Yes | Advertised |
| Cost | $10/mo ($120/yr) billed annually; free Starter tier | $15/mo, or $12/mo billed annually; former AppSumo deal sold out |
| Free tier | 2,000 lifetime words | 1,000 words per month |
| Platforms | Mac, Windows, iPhone, Android | Mac, Windows, Linux, iPhone, Android |
| Languages | 49 | 50+ (self-reported) |
| Works offline | No (cloud app, needs internet) | No (cloud-only) |
| Processing & security | Managed cloud; SOC 2 Type II audit report, ISO 27001, GDPR, HIPAA | Managed cloud; Zero Data Retention, BAA with provider |
| Developer API | Avalon API, $0.39/hr, OpenAI-compatible | Transcription API on plans that include it |
| Best for | Accuracy-first users on Mac, Windows, iPhone, Android | Linux users; recurring free allowance |
A model Aqua actually owns
Avalon is built and trained in-house, not rented from a third-party API, and scores 97.4% on AISpeak-10. VoiceDash publishes no independent accuracy number.
Screen-context awareness
Aqua reads your screen in real time to get names, code, and jargon right. VoiceDash adapts formatting to your app, but does not describe reading on-screen names or code.
Lower subscription price
Aqua Pro is $10/month billed annually. VoiceDash Pro is $12/month billed annually, or $15 month to month.
Linux
VoiceDash runs on Linux as well as Mac, Windows, iPhone, and Android. Aqua Voice has no Linux app.
A monthly free tier
You get 1,000 free words every month. Aqua's free tier is 2,000 words in total for the life of the account.
More languages
VoiceDash lists 50+ languages, by its own count. Aqua Voice supports 49.
Key AI and coding terms written right on AISpeak-10, Aqua’s English test.How we measuredOpenASR leaderboard
VoiceDash cites an "OpenAI partnership" but does not name the speech model it runs. It publishes no accuracy benchmark, only a "98% user satisfaction rate," so we cannot compare it with Avalon.
VoiceDash launched in 2025. It sends speech to a third-party AI provider, then removes filler words and adds punctuation. This lets it ship on five platforms quickly. Source: VoiceDash product information, checked September 30, 2026.
Aqua trains its own model, Avalon, and links it to real-time screen context. That focus costs Linux support, but it lets Aqua publish an accuracy number.
VoiceDash publicly says it processes speech through an "OpenAI partnership" with Zero Data Retention and a signed BAA, but it does not name the specific model it runs, and we won't guess at one. Aqua is different: it uses Avalon, its own model built and trained in-house, which is why Aqua can publish its own benchmark.
Aqua publishes 97.4% accuracy on the AISpeak-10 benchmark and ranked #1 among proprietary systems at its OpenASR debut in October 2025. VoiceDash publishes no independent accuracy benchmark, only a '98% user satisfaction' figure, so there is no apples-to-apples number to compare. On measured accuracy, Aqua is the one with the receipts.
On subscription, no. Aqua Pro is $10/month billed annually ($120/year) versus VoiceDash Pro at $12/month billed annually ($15 month to month). VoiceDash's edge is a free tier that resets every month (1,000 words every month vs Aqua's one-time 2,000), while its former AppSumo lifetime deal is sold out.
Android, yes; Linux, no. Aqua supports Mac, Windows, iPhone, and Android. If you need Linux, VoiceDash covers it and is the better fit there.
No. Both Aqua and VoiceDash are cloud apps that need an internet connection to transcribe. If you specifically need offline, on-device dictation, neither is the right tool.
Aqua reads what is on your screen in real time, so it formats correctly and gets on-screen names, code, and jargon right instead of transcribing audio blind. VoiceDash adapts formatting to the app you are in, but does not describe reading on-screen names or code.
Aqua offers the Avalon developer API at $0.39/hour, and it is OpenAI-compatible, so you can build on the same model that powers the app. VoiceDash added a transcription API in September 2026, on plans that include API access, with usage counted against the plan's word quota.
Used by people at
Students get 70% off Pro and Max.