Dictation tools, ranked by what real users say.
The Dictation Index scores voice dictation software on aggregated public user sentiment instead of one reviewer's opinion. Every score is published with its sample size, its date range and the method that produced it.
Every ranked dictation tool, by user sentiment.
All 12 tools with at least 50 qualifying posts, scored on aggregated public sentiment. Each row shows its margin of error, the dates its posts cover and the ranks it is statistically tied with.
Bayesian sentiment score · 90% margin per row
-
#1
Aqua VoiceHighest score
Cloud AI dictation with low latency and custom dictionary, for Mac, Windows and iOS
CloudmacOSWindowsiOSFreemiumReal-timeCustom vocabContext-aware7.1 / 10n = 606 -
#2
SpokenlyTop on accuracyTop on price
Privacy-first Mac and iOS dictation with free local models and optional cloud AI
HybridmacOSiOSFreemiumOfflineSystem-wideAPI7.0 / 10n = 244 -
#3
HandyTop on privacyTop on platform supportTop on usability
Open-source offline dictation for Mac, Windows and Linux, MIT licensed
On-devicemacOSWindowsLinuxFreeOfflineOpen sourceSystem-wide7.0 / 10n = 176 -
#4
VoiceOS
Context-aware dictation with an agent mode that acts on apps, for Mac and Windows
CloudmacOSWindowsFreemiumContext-awareVoice cmdsSystem-wide6.9 / 10n = 80 -
#5
OpenWhisprOpen source
Open-source Wispr Flow alternative with local Whisper, MIT licensed
HybridmacOSWindowsLinuxiOSFreeOfflineOpen sourceSystem-wide6.9 / 10n = 270 -
#6
Willow Voice
Context-aware AI dictation with custom vocabulary, optional HIPAA compliance
HybridmacOSWindowsiOSAndroidFreemiumContext-awareCustom vocabSystem-wide6.8 / 10n = 166 -
#7
VoiceInkTop on speedOpen source
On-device Whisper and Parakeet dictation for Mac, one-time purchase, GPL licensed
On-devicemacOSOne-timeOfflineOpen sourceSystem-wide6.7 / 10n = 169 -
#8
MacWhisperOn-device
Local Whisper dictation and file transcription for Mac, one-time purchase
On-devicemacOSOne-timeOfflineDiarizationMulti-lang6.6 / 10n = 145 -
#9
Superwhisper
Local Whisper dictation for Mac and iPhone, with optional cloud models
HybridmacOSiOSSubscription or one-timeOfflineSystem-wideAPI6.0 / 10n = 554 -
#10
TypelessMost discussed
AI dictation with filler removal and tone adaptation, cross-platform
CloudmacOSWindowsiOSAndroidFreemiumReal-timeContext-awareFormatting5.6 / 10n = 769 -
#11
Wispr Flow
Cloud AI dictation that types into any app, for Mac, Windows, iOS and Android
CloudmacOSWindowsiOSAndroidFreemiumReal-timeContext-awareSystem-wide5.6 / 10n = 668 -
#12
Apple Dictation
Built-in dictation on Mac, iPhone and iPad, on-device on Apple Silicon
HybridmacOSiOSFreeReal-timeMulti-langAuto-punct3.2 / 10n = 120
57 more tools are in the registry without a rank, because too few qualifying posts mention them. See all 69 tools in the registry →
A ranking that shows its work.
Most "best dictation software" pages rank on one writer's opinion and publish no sample size. The Dictation Index counts public posts, prints the sample size and margin beside every score, and shows each tool's losses next to its wins.
The formula is fixed and public
Every post counts 10, 5 or 0. The prior, the minimum sample, the filters and the method version are on the method page.
No paid placement
No vendor can buy a position, a badge, a review or a mention. Listing is automatic for anything that meets the definition.
Losses are published
Every ranked tool's page shows where it loses, topic by topic. Even on the highest-scoring tool's page that means reliability (3 positive posts against 32 negative) and price (30 positive posts against 44 negative).
What you cannot check yet
The post IDs behind each score are not public yet, so the counts cannot be re-run independently today. Until they are, read every score as our count, shown with its sample size and margin. Known limits of the method.
How a score is made
The Index replaces the reviewer with the crowd, then shows its work. Four steps, each of them published. The full method adds the statistics and the known limits.
Collect
Public posts on X that mention a tool by name, from the past 90 days. Collection currently reaches back up to 1,000 posts per tool, so the busiest tools cover a shorter span. Every ranking row and tool page states the dates its posts cover.
Classify
Each post is labelled for sentiment (positive, negative, neutral or mixed) and for topic: accuracy, speed, price, reliability, privacy, platform support, support and usability.
Filter
Official vendor accounts, automated reply bots, duplicate posts and posts that could not be labelled are removed. No single author can contribute more than three posts to one tool.
Score
Each post counts 10 if positive, 5 if neutral or mixed, 0 if negative. The average is shrunk toward the scale midpoint by a sample-size prior, so twenty glowing posts cannot outrank two thousand mostly positive ones.
score = (10 × pos + 5 × neu + m × μ) / (n + m)
Pre-filled with the current counts for Aqua Voice.
- pos
- qualifying posts labelled positive (10 points each)
- neu
- posts labelled neutral or mixed (5 points each); negative posts score 0
- n
- all qualifying posts for the tool in the window
- μ
- the prior mean, fixed at 5, the midpoint of the scale
- m
- the prior weight, fixed at 5 posts per method version
Published with every score:
- n, the margin of error and any statistical tie
- the dates the tool's posts actually cover
- the topic breakdown, so you can see where a tool wins and where it loses
- the method version
Tools below the minimum n are listed but not ranked.