murmr
198 posts
murmr
@murmrvoice
The app to talk and listen to your computer. Your transcripts never leave your device. MacOS, Windows, iOS and Android (soon).
Katılım Haziran 2026
104 Takip Edilen34 Takipçiler
making a gc for hackathon builders
reply to be added
English
@murmrvoice I did. It listens but does nothing after listening. Just nothing happens!
English
62ms response time. Private. Local. Free.
Alright, I'm speaking this tweet using the latest model that we've refined, performance and accuracy-wise, so that this dictates instantly with an average response time of sixty-two milliseconds. This model is the fastest model available right now on Mac and Windows that runs locally and privately. It's available on our website for direct download to Mac and Windows will be released on their respective stores here in the next couple of days.
English
murmr retweetledi
FluidVoice is officially replacing WisprFlow for me.
If you have not tried this kind of utility, it lets you talk to your machine instead of typing. Voice straight into prompts, docs, anywhere text goes. The productivity boost is big for vibe development. It tremendously helps even if you are good at typing.
WisprFlow easily 20x my vibe coding. But it costs $20 a month and runs in the cloud, so my voice data leaves my machine.
FluidVoice is open source and completely free. It runs fully local, nothing leaks to the cloud. And honestly, in my opinion, it is better.
If you want to 10x your vibe coding productivity, give it a try. altic.dev/fluid
#BuildOnXRPL #XRP #XRPL
English
@daniell0930 @ordonez_adan That's an interesting perspective. What about a dictation service that uses local models only? Transcripts / voice don't leave the device.
English
@ordonez_adan I’m a litigator so this may be specific to us. But litigators are always nervous & concerned abt attorney work product & privilege. I think that’s why you see an under utilization of some of these tools. Bc what are the terms of Wispr Flow, can it get out of my sphere..
English
I've yet to meet a lawyer taking advantage of AI voice dictation. The gains in work speed are insane.
Tools like Wispr Flow turn your voice into text - nothing crazy - except that the output is quite accurate and allows you to direct various AI agents in a short amount of time to complete various tasks.
Instead of typing a long prompt, you speak to it, the same way you would walk an associate through something out loud. For example, you could tell your AI tool to research or draft a memo summarizing its findings. You can ramble on for minutes and get very specific. The tool will pick up each detail, unlike a human, who may forget.
This is the future.
English
No one pays for dictation anymore bro
Willow@WillowVoiceAI
Why are you paying for dictation? We're releasing free, unlimited AI dictation. And it's not a slow, local model. Willow Frontier Mini is cloud-based with zero-data-retention. More accurate and faster than Wispr Flow, OpenAI, Deepgram, and more. See video comparison.
English
@amaan8429 What information are you dictating? x.com/TheThomHart/st…
Thom@TheThomHart
I was a Wispr Flow power user. 1M+ words dictated, relied on it religiously - did my morning thoughts on Wispr Flow, used those transcripts to ideate for the day. Then I started thinking about how much of my life was going on Wispr Flow; their servers, my transcripts, my voice - my life. Then, in early June, their service started degrading. It became unreliable and spotty. So I built my own. The idea was that it would always be private. All transcripts were local and never left the phone. No voice capture off device. No servers. All maximally private. Decided I would add all the features I wanted: speech to text, text to speech, meeting recordings, terminal integrations via MCP, etc. Many more features to come. Now it's live on iOS and Mac. And Pro is free for builders, because I care more about connecting and learning from them than charging for this service. I'm continuing to make the model faster and more accurate, and have no doubt it will be on par with Wispr Flow very soon. Check it out at murmrvoice.com / @murmrvoice and let me know what sucks. I'll fix it same day.
English
Ran out of my Wispr Flow limits for the week (free tier).
Asked Codex to study the Wispr Flow desktop app and build an alternative using free STT models from Groq or other providers.
GPT-5.6 Sol Extra High spent 18 minutes planning and exploring Wispr Flow using computer use.
GPT-5.6 Sol Medium spent 25 minutes implementing it.
In the end, I had a fully functional app I could actually use.
@OpenAI Codex is a beast.
English
@alexzfunk @WisprFlow Alright that's pretty good. I don't think we can compete with that lol
English
@WisprFlow is a perfect product - blasting music while working and it just understands perfectly
English
@olawaleflair Doesn't work on all applications. Apps like murmr and wispr work on all text entry fields. murmr also works offline. Codex doesn't
English
@themarketradar Agree. That's why I created my own local dictation / text to speech app.
English
It’s really insane how garbage apple dictation is compared to the dictation options on the Claude and GPT apps holy shit
English
@kevinmanase @WillowVoiceAI How do you like it? Mac or iOS? We just pushed a great update to Mac if you want to redownload directly from the site
English
Why are you paying for dictation?
We're releasing free, unlimited AI dictation.
And it's not a slow, local model.
Willow Frontier Mini is cloud-based with zero-data-retention. More accurate and faster than Wispr Flow, OpenAI, Deepgram, and more.
See video comparison.
English
Windows beta is live now and we'll be released on the windows store shortly. Enjoy!
murmrvoice.com/download
English