How I build apps these days: 1. Record a 1-5 minute prompt with @WisprFlow The more detail, the better. 2. Add visual inspo so the agent knows what I have in mind (screenshots or Inspo MCP). 3. Use @mattpocockuk's grill-me skill to have the agent ask clarifying questions. 4. Get back a v1, then send a list of feedback, feature requests, and design tweaks (via another voice recording usually). 5. For each new feature, ask for a quick screen recording to verify it works before merging the PR. 6. Keep iterating until I'm happy, then post it on @X or throw it away. For my harness, I use a mix of Claude Code with Opus 5.5 (high) and OpenCode with DeepSeek V4.1 Flash / GLM 5.3. I also give my agents some CLIs & skills: • Vercel + GitHub CLIs so they can push code and deploy apps. • Together skills so they can use chat, image, & voice models from Together • Design/animation skills so they have guidance on spacing & interactions
Wispr Flow on X
The talk about Wispr Flow is overwhelmingly positive: 145 of 286 people in the reception posts recommend it without caveats. What repeats in the complaints is reliability and the phone app, and the people who leave go to free, local dictation they build or install themselves.
First written 27 Oct · what changed · from 322 posts by 286 people
Tracking posts that say “@wisprflow” or “wispr flow”
Track this for your marketPosts per day, last 30 days
as of Nov 3, 09:00 UTCWhat changed
as of 27 Oct · checked once a monthFirst written 27 Oct. When what people say changes, it shows here.
Tone· 199 posts from 182 people· written 27 Oct
How do people feel about Wispr Flow?
145 of the 182 people in this section recommend it without caveats: the best dictation they have used, worth telling others about. 19 like it but name a frustration in the same breath, 14 say it has got worse or are leaving, and 5 are still deciding.
Enthusiasts describe a tool they use all day: dictating prompts to Claude Code and Codex, emails, meeting notes, messages and posts, in French, Portuguese, Chinese and Hindi as well as English. They count what they have spoken — a quarter of a million words, a 132,550-word streak at 116 words a minute — and say it multiplied their output; one calls it "goated for working with AI", another taught a parent to use it in an afternoon and found that one button was all it took.
Reservations are specific: unwanted quotation marks in texts, autocorrect below Apple's, an iPhone app that loads like another keyboard, context they have to keep restating. The disappointed say it fails several times a day or gets worse the longer they use it, and some of those verdicts are about price and a $2bn valuation rather than the product. The undecided are still trying it or weighing a subscription; one came because the iPhone's own voice-to-text is bad.
- Enthusiastic161
- Satisfied with reservations19
- Disappointed14
- Undecided5
See 5 of 199 postsHide the posts
Officially hit a quarter-of-a-million words dictated by @WisprFlow. I think it’s the most used tool in my workflow on a daily basis. 4x’d my input, 10x’d my output. I can’t recommend it more.
I taught my dad to use @WisprFlow in one afternoon. He loved it. One button. It just works. Too many tech founders debate inference and the latest MCPs on X, then build for their own curiosity, not paying customers. What’s the last tool you got a parent to actually use?
Wispr Flow has great marketing + I love the user dashboard but dealing with this several times per day is a complete deal breaker. Every single LLM default dictation works far more reliably than this.
The more I use @WisprFlow - the worse it gets. Is it just me? In the morning works really good. Closer to the evening it is worse than apple keyboard built in. Seriously considering cancelling my sub and vibecoding an alternative
5 of the 199 posts behind this answer, one per person. The other 194, and every new one as it’s posted, come with tracking. Get early access
Pain points· 68 posts from 62 people· written 27 Oct
What do people complain about with Wispr Flow?
Reliability is the biggest complaint: 13 people describe dictation that fails, gets stuck or misses the text box, and 11 say it mishears words. The phone app (9 people) and language output that flips (7) come next, and several of them say it worked before and has got worse.
Failure to work at all leads. Users name Chrome text boxes not being picked up, text that will not paste, a black screen on opening the app, a mic not detected on one Windows 11 build, and dictation that gets stuck in web chat boxes. A self-described top 10% user asks for a fix; another calls it unusable, and several say it worked before and has got worse — one finds it worse than the built-in Apple keyboard by evening.
Mishearing follows: a sneeze turning done into duck, filler words like wait and okay so dropped, names and technical terms mangled, quotation marks appearing in texts, autocorrect below Apple's.
Phone complaints are concrete: no globe key to switch back, no swipe typing, an awkward "switch to the Wispr Flow app temporarily and then go back", a persistent Android icon, an iPad app to fix. Language output flips from Hinglish into Devanagari or Urdu, and an English call was written in Telugu.
Price, word allowances on paid plans, cloud audio, nearby conversations, memory use and no Linux build are each raised by a few.
- Unreliable dictation14
- Misheard words12
- Mobile app shortcomings9
- Wrong language output7
- Subscription price5
- Word limits on paid plans4
See 5 of 68 postsHide the posts
@WisprFlow as a top 10% user, there is something fundamentally going wrong with Wispr Flow in Chrome. It is not picking up the text boxes properly, and I have had to waste a lot of time finagling with Wispr Flow and Ctrl+Cmd+V's because it is not properly pasting text. Please fix this ASAP.
Wispr Flow has great marketing + I love the user dashboard but dealing with this several times per day is a complete deal breaker. Every single LLM default dictation works far more reliably than this.
The more I use @WisprFlow - the worse it gets. Is it just me? In the morning works really good. Closer to the evening it is worse than apple keyboard built in. Seriously considering cancelling my sub and vibecoding an alternative
No, no. I actually wanted to say, "too", that "he is an actual developer, too". @WisprFlow Why did you do me like that? 😂
I can't wait for the day when Wispr Flow on iOS doesn't need to do that awkward "switch to the Wispr Flow app temporarily and then go back to the app you're on" to work...
5 of the 68 posts behind this answer, one per person. The other 63, and every new one as it’s posted, come with tracking. Get early access
Alternatives· 57 posts from 55 people· written 27 Oct
What are the alternatives to Wispr Flow?
Building a local setup yourself (8 people) and Superwhisper (8) are level at the top of what people weigh it against, with Apple's built-in dictation and the mic inside ChatGPT, Claude and Codex just behind at 7 and 7. The pull is towards free or local, and the argument is as much about price and privacy as accuracy.
People who build their own describe replacing the subscription in a couple of prompts or twenty minutes with a local Mac app, sometimes on Groq or DeepSeek APIs, at a dollar or two a month, and say the result is close enough. Superwhisper splits both ways: some call Flow weak next to it or switch, others tried it and came back.
The free built-ins are the other test. A Mac already has dictation, one says; another broke a 258-day Flow streak because ChatGPT's dictation on Windows was that good; others now dictate inside Claude desktop and mobile, or paste from ChatGPT into Claude Code.
FluidVoice is the free local app most often named as the replacement, for switching models, running locally and privacy — though some say it is not as fast or smooth. Kiwi by Sarvam is named for Indian accents and Hinglish. Granola is the meeting-notes comparison, with one saying Flow's notetaker makes it duplicative and another preferring Granola. Google's free local tool, Willow, Muse, Handy, Sayso, Spokenly, Typeless and Raycast appear once or twice each.
- Built it myself9
- Superwhisper8
- Apple built-in dictation7
- Built-in AI app dictation7
- FluidVoice6
- Granola4
See 5 of 57 postsHide the posts
Retour sur mon wispr flow maison: - très rapide - feature dictionnaire qui fonctionne très bien - aucun soucis de focus J'ai aucune envie de repasser sur Wispr Flow en payant 15$/mois, c'est tellement plus rentable mon app perso.
I just replaced my Wispr Flow subscription in 2 prompts. I now have a penalized mac app that works 100% locally and I can customize it anyway I want. Took me 20 minutes to make…and yes I asked Claude to automatically check and update the app if needed each week. So it can download newer open source models each time they become available.
it was quite good. I'm embarrassed to say I mostly use @WisprFlow. I try to replace it every few months and fail (Superwisper, for instance, is weak AF in comparison). I do dictate a lot.
wispr flow is shit @superwhisper mogs
I broke a 258-day Wispr Flow streak because ChatGPT's dictation on Windows is that good and that fast. But a few things would make it a no-brainer. Here's my feedback after using both side by side. The bug first: If I start dictating inside a ChatGPT chat and switch chats, I lose everything and have to start over. If I start dictation outside the app, it survives the switch. Dictation should persist no matter where it started. Shortcuts and control 1. Flexible shortcuts. Let me use Alt+Q for hands-free mode, and assign more than one shortcut. 2. A separate cancel key. Esc cancels dictation AND stops the AI response. I want to cancel one without killing the other. 3. "Paste last transcript." If a paste fails, let me recover it instantly. Editing power 4. Prompt-based transforms. Select text, hit a shortcut, apply a saved prompt (like Wispr Flow's Transform). 5. Polish levels: none, light, medium, high. Sometimes I only want spelling and punctuation fixed, not my voice rewritten. 6. App-specific styles. WhatsApp shouldn't sound like LinkedIn. Same for prompts and notes. Vocabulary 7. Word replacements. Fix a name it keeps getting wrong once, forever. 8. Better dictionary management, with imports from tools like Wispr Flow. 9. Agent-suggested vocabulary, added automatically with my permission. 10. Spoken snippets. "My email address" expands to my actual email. The big ideas 11. Usage analytics for dictation. 12. Learn my natural voice. Dictation captures how I actually talk. Turn that into an editable writing-style profile my agent can use. If I had to pick three: fast text transforms, dictation that survives chat switches, and control over how much my words get changed. What would you add? cc @athyuttamre @thsottiaux @gdb @OpenAI
5 of the 57 posts behind this answer, one per person. The other 52, and every new one as it’s posted, come with tracking. Get early access
Questions· 22 posts from 19 people· written 27 Oct
What do people ask about Wispr Flow?
The largest group, 3 people, wants a verdict from someone who has used it: how do you like it, and where does it fit. The rest of these 19 people ask single, practical things — Linux, phones, a local option, your own key — and no other question is asked by more than two.
The ask that repeats is a verdict rather than a feature. People ask what others think of it, whether it is better for long sessions in one app or for switching across email, charts and X, and — of a user who has dictated 132,000 words — whether it still trips on names and technical terms.
Practical questions are about fit: whether it works in a particular app, whether you can bring your own API key or local model instead of a subscription, whether there is a Linux version or a Linux alternative, and whether it ignores them talking to themselves. One asks for good local options because they are uneasy about how much leaves the device, screen context included.
There are single asks too: how much the cleanup and beautification adds over sending the raw ramble, whether a long spoken prompt follows the start or the end when the speaker changes their mind, who uses it at all, whether Apple has bought it yet so they can dictate on a phone again, and how to get in through the Trybe creator programme.
- How do people like it?4
- Is there a free alternative?2
- Is there a Linux version?2
- Does it filter out self-talk?2
- Can I use it on my phone?2
- Who is using it?2
See 5 of 22 postsHide the posts
So what are people's thoughts on Wispr Flow? (I meant "deck")
What do people use for AI-powered voice dictation? I like Wispr Flow but I get creeped out by how much it sends off to them (including screen context). Are there good local options?
132K words is serious. Does it still trip on names and technical terms, or is it clean by now?
Any good Linux alternatives for Wispr Flow?
Can you use your own API key or local model? Or does it require a subscription?
5 of the 22 posts behind this answer, one per person. The other 17, and every new one as it’s posted, come with tracking. Get early access
Wishes· 26 posts from 25 people· written 27 Oct
What do people want from Wispr Flow?
No wish is big: a Linux app (3 people), a better mobile keyboard (3), notes always in English (2) and customisable shortcuts (2) are the largest of what these 25 people ask for. Together they want the phone to work properly and more control over what comes out.
Linux is the most repeated ask, and it is framed as lost business: one user went back to Windows 11 to keep using Flow, and another calls for a competent Linux alternative because "1000s of us are ready to pay monthly".
The phone comes next, overlapping with the complaints: an updated iPad keyboard, an iPhone keyboard that is only a transcription button, swipe typing, and a globe key to get back to another keyboard.
Then control over what comes out. People want notes always in English even when they speak Hinglish, Hinglish in Latin script rather than Devanagari, and dictation that handles English and Chinese in one go; modes without punctuation or capitals; the shortcut turned off entirely, several shortcuts, and a cancel key separate from the AI's stop.
Smaller asks: seeing text as they speak rather than at the end, pausing audio while dictating, folders for meeting notes, moving data between work and personal accounts, a walking-around mode with just mute and edit, remembering project context, and a daily page of the mistakes the cleanup fixed.
- A Linux app3
- A better mobile keyboard3
- Notes always in English2
- Customisable shortcuts2
- Notetaker for iOS2
- Live text while speaking1
See 5 of 26 postsHide the posts
Can someone make a competent wispr flow alternative for Linux please? 1000s of us are ready to pay monthly. Had to go back to windows 11 for Wispr flow on my personal laptop.
@WisprFlow two requests: 1. Release an iPhone Duo keyboard where it's just a transcription button and not a keyboard. 2. Please fix your keyboard. There's no swipe-to-type. Just buy a keyboard app from the App Store for like a million bucks.
@WisprFlow team - absolutely loving the product! Small feature request: Can we have an option to always generate notes in English, even when speaking in Hinglish? Right now, it switches between English and Hinglish inconsistently.
I broke a 258-day Wispr Flow streak because ChatGPT's dictation on Windows is that good and that fast. But a few things would make it a no-brainer. Here's my feedback after using both side by side. The bug first: If I start dictating inside a ChatGPT chat and switch chats, I lose everything and have to start over. If I start dictation outside the app, it survives the switch. Dictation should persist no matter where it started. Shortcuts and control 1. Flexible shortcuts. Let me use Alt+Q for hands-free mode, and assign more than one shortcut. 2. A separate cancel key. Esc cancels dictation AND stops the AI response. I want to cancel one without killing the other. 3. "Paste last transcript." If a paste fails, let me recover it instantly. Editing power 4. Prompt-based transforms. Select text, hit a shortcut, apply a saved prompt (like Wispr Flow's Transform). 5. Polish levels: none, light, medium, high. Sometimes I only want spelling and punctuation fixed, not my voice rewritten. 6. App-specific styles. WhatsApp shouldn't sound like LinkedIn. Same for prompts and notes. Vocabulary 7. Word replacements. Fix a name it keeps getting wrong once, forever. 8. Better dictionary management, with imports from tools like Wispr Flow. 9. Agent-suggested vocabulary, added automatically with my permission. 10. Spoken snippets. "My email address" expands to my actual email. The big ideas 11. Usage analytics for dictation. 12. Learn my natural voice. Dictation captures how I actually talk. Turn that into an editable writing-style profile my agent can use. If I had to pick three: fast text transforms, dictation that survives chat switches, and control over how much my words get changed. What would you add? cc @athyuttamre @thsottiaux @gdb @OpenAI
@WisprFlow is super cool, but it would be cooler if you could actually see the text rendering as you speak instead of all at once at the end (I've been yapping for a very long time and then, in the end, realized that actually, I wasn't recording anything)
5 of the 26 posts behind this answer, one per person. The other 21, and every new one as it’s posted, come with tracking. Get early access
In their words
How do people talk about Wispr Flow?
The phrases people use most, counted once per person.
- 1“voice dictation”6
- 2“open source”6
- 3“voice typing”4
- 4“please fix”3
- 5“worse than apple”2
- 6“speech to text”2
- 7“free and open”2
- 8“rewriting and organizing”2
- 9“words dictated”2
- 10“dictation tool”2
Posts are shown the way X shows them. One of them yours? Here’s why it’s shown, and how to take it off.
Also in Voice dictation
- Superwhisper—Talk about Superwhisper is mostly people recommending it by name for everyday dictation and citing the lifetime sub or free tier as why they stay.226 posts / 7d
- Aqua Voice−65%Talk about Aqua Voice is praise-led: users say it writes down what they meant, fillers stripped, with the pixel pet and its leaderboard the next strand.141 posts / 7d
- Willow−85%The largest theme is praise: 23 people recommend Willow and 9 single out Wally, its writing agent.93 posts / 7d
- Monologue—Talk about Monologue is calm and positive: the largest group, 10 people, are low-key satisfied users who dictate daily.20 posts / 7d
Get every post about Wispr Flow, filtered and labelled
This page is the summary. Tracking gives you the stream behind it for any brand, keyword or @mention: each new pain point, request and question as it’s posted, with the spam and promotions already taken out.





















