Aqua Voice on X

Praise is mostly about accuracy in the users' own sense: people say it gets down what they meant, strips the fillers, and keeps their phrasing. One heavy user says it extracts the intent in their voice without tidying too much or too little; others point to mixed Japanese-and-English sentences, punctuation and paragraphing, and a personal dictionary that carries across phone and desktop. Several count their words in the hundreds of thousands.

The Android release drew its own run of praise, largely because dictating needs no keyboard switch: a tap on the floating button types into whatever field is open.

The pet is the second strand (17 people): a pixel companion that grows as you dictate, with unlockable outfits, streaks and a leaderboard ranking users against each other. Fans call it gamification done well and post their word counts and streaks; one long-time user calls the pet cute but says the other updates matter more.

Smaller groups: 5 say dictating beats typing and they can't go back; 4 have cooled, finding the output short of rivals or the built-in tools close enough; 3 weigh paying once the free allowance runs out.

The pet is the biggest single complaint (6 people). One turned it off because it was too big and distracting and asked for a resize option; another says that on Windows it takes the foreground and breaks background runs; another cannot find the pet in the latest version; one had an unfamiliar mascot appear unbidden. One who shrank it calls it tolerable.

Slowness and failures come next (5 people): one says dictation often produces no text at all and has to be copied out of the app and pasted, one reports crashes, another consistent slowness on Windows before about 10 a.m. JST, one lost a long dictation until it turned up in history, and one sees delays starting on iPhone.

4 people are unhappy with the output — Japanese accuracy, new proper nouns, non-English words, and quality short of ChatGPT. The mic wake delay is two people, level with the 2 who won't pay for dictation: worse over Bluetooth and on a Studio Display, helped by keeping the mic open, which one says forces mono.

Built-in dictation leads (6 people), in both directions. Some find Aqua better than Apple's: one says it is much more accurate than the iPhone's own dictation and another would rather use it than the stock option, while a third recommends it alongside the Mac default. Others find the built-in good enough — one who quit says Claude Code's own dictation is no big difference, and one weighing the free allowance asks whether they need it beyond the mic that AI tools already have.

Typeless is next (5 people). Android users like that Aqua needs no keyboard switch, while those on Typeless cite its monthly free allowance and automatic tidying of self-corrections; one switches to Typeless anyway because Aqua mishears new proper nouns, and one returns to Aqua because Typeless over-structures text and misses proper nouns.

WisprFlow is two people: one prefers Aqua on technical terms but stays with Wispr for non-English languages, another goes back and forth between the two. Dots also two: time talking to Dots leaves less reason to dictate. The rest are single mentions — Voiceboard, a Whisper build of their own, Rambler, Monologue, Handy with Parakeet, Letterly, Simeji, Flow, FluidVoice.

With 2 people asking, the questions are narrow. One asks straight whether it works in French. The other, posting in English, asks whether it only works in Japanese or understands English voice too — a question about what language the tool takes, not about translation.

Neither post carries an answer, and nothing filed under this section reports back on either. Elsewhere on the page the language question shows up in what people say rather than ask: users describe mixed Japanese-and-English dictation as a strength, and someone who compared it with WisprFlow says the other tool handled Latvian much better while this one was stronger on technical terms.

For anyone choosing between dictation tools, the section says the public questions here are about coverage before anything else — which languages are supported, and whether English is one of them alongside Japanese.

No wish is shared by more than one person, so none of them is a trend; they are three separate asks.

The first comes from someone who likes the new pet animation and wants "that same level of polish" to reach mobile devices. The second is a repeat request for a "most words in a day stat" — the same person had asked before, and the leaderboard and streaks they already see are about week-to-week word counts, not a daily best. The third asks whether the pill can be moved on macOS, which is about where the control sits on screen rather than how dictation works.

Two of the three are about the game layer, the leaderboard and the pet, rather than about transcription itself. The section is thin at 3 people, so read it as three individual notes rather than what users in general want.