Most voice typing lives in the cloud, costs a monthly fee, and eats your GPU. Levitan doesn't. Hold a key, speak, and polished text appears where your cursor is - in any app. Runs locally on your CPU, fully offline, nothing leaves your device. Free forever. RU/KZ/EN + 96 languages. Windows & macOS.
No reviews yetBe the first to leave a review for Levitan
Maker
📌
Hey Product Hunt 👋
I'm Konstantin, founder of Loqira Labs. For the past months I've been building Levitan — a voice-to-text app that runs entirely on your own computer.
The itch I kept scratching: every good dictation tool sends your voice to the cloud, charges a monthly fee, and still makes the fan spin. I wanted the opposite — something that just turns speech into text, locally, for free, and gets out of the way.
So with Levitan you hold a key, speak, release — and the text appears right where your cursor is, in any app: chat, email, your code editor, an AI prompt.
What makes it different:
🔒 On-device & private — speech is processed on your CPU. Nothing goes to the cloud. Not "we promise not to look" — we technically can't.
🪶 Light — no GPU, no gaming laptop. It runs on an ordinary CPU and stays out of your way.
🆓 Free, no asterisks — no subscription, no minute or word limits. Voice input shouldn't be another monthly bill.
🌍 Multilingual — Russian, Kazakh, English + 96 more.
⚙️ Our own engine — native Rust, a quantized & finetuned Whisper model, custom CPU inference. Not a wrapper over someone's API.
It's especially handy for long AI prompts — when a good task for ChatGPT / Claude Code / Cursor doesn't fit on one line, it's just faster to say it out loud.
Available today on Windows and macOS (Apple Silicon). Android & iOS are next.
I'd genuinely love your honest feedback — what feels off, what's missing, which language or accent I should improve next. I'll be here all day answering everything. 🙏
— Konstantin · loqira.tech
Report
Privacy-first AI is one of those spaces where trust can become a much stronger moat than features.
Curious—do users adopt Levitan because it's local/private, or do they stay because the workflow ends up being better than cloud alternatives?
Report
Maker
@aryan787544 It depends on the user’s needs. For those who simply need dictation, Levitan fully meets their requirements, so they continue using it. It is especially useful when working on a laptop somewhere without an internet connection and dictating thoughts, notes, or ideas.
I'm curious whether that changes over time as well. Someone might adopt it because it's private and offline, but still end up relying on it for a completely different reason months later. It'll be interesting to see if those reasons diverge as your user base grows.
Report
Privacy-first AI is becoming a much more interesting space than people expected.
Curious—what's been the bigger driver of adoption so far: privacy, or the experience once people actually try Levitan?
Report
Maker
@aryan787544 It depends on the user’s needs. For those who simply need dictation, Levitan fully meets their requirements, so they continue using it. It is especially useful when working on a laptop somewhere without an internet connection and dictating thoughts, notes, or ideas.
Report
This is genuinely cool, offline voice typing that just works sounds amazing. One thing that would make it a daily driver for me is a custom vocabulary or glossary feature, letting me add jargon, names, or technical terms so it stops mangling them in transcripts.
Report
Maker
@eyupsapc10102 I’m planning to add this together with a small local LLM that will handle punctuation, paragraph breaks, and dictionary-based word correction.
Report
the offline angle is genuinely cool and the free model makes this way more approachable than whisper-based alternatives i've tried. one thing that would actually push me to use it daily though, could you add a custom vocabulary or replacement rule system? like so i can set "levitan" to always become "Levitan" or auto-format certain phrases i say a lot. would make the raw transcription feel way more polished without me having to fix stuff every time.
Report
Maker
@kocanaoglu80506 I’m planning to add this together with a small local LLM that will handle punctuation, paragraph breaks, and dictionary-based word correction.
Report
honestly love the offline angle, super rare these days. one thing i'd actually use though is a custom vocabulary or shortcuts feature so i can teach it my weird jargon and have it auto-replace spoken phrases with longer snippets, that would save me way more time than just raw dictation
Report
Maker
@sendurgut14824 I’m planning to add this together with a small local LLM that will handle punctuation, paragraph breaks, and dictionary-based word correction.
Privacy-first AI is one of those spaces where trust can become a much stronger moat than features.
Curious—do users adopt Levitan because it's local/private, or do they stay because the workflow ends up being better than cloud alternatives?
@aryan787544 It depends on the user’s needs. For those who simply need dictation, Levitan fully meets their requirements, so they continue using it. It is especially useful when working on a laptop somewhere without an internet connection and dictating thoughts, notes, or ideas.
@kolomeitsev That makes sense.
I'm curious whether that changes over time as well. Someone might adopt it because it's private and offline, but still end up relying on it for a completely different reason months later. It'll be interesting to see if those reasons diverge as your user base grows.
Privacy-first AI is becoming a much more interesting space than people expected.
Curious—what's been the bigger driver of adoption so far: privacy, or the experience once people actually try Levitan?
@aryan787544 It depends on the user’s needs. For those who simply need dictation, Levitan fully meets their requirements, so they continue using it. It is especially useful when working on a laptop somewhere without an internet connection and dictating thoughts, notes, or ideas.
This is genuinely cool, offline voice typing that just works sounds amazing. One thing that would make it a daily driver for me is a custom vocabulary or glossary feature, letting me add jargon, names, or technical terms so it stops mangling them in transcripts.
@eyupsapc10102 I’m planning to add this together with a small local LLM that will handle punctuation, paragraph breaks, and dictionary-based word correction.
the offline angle is genuinely cool and the free model makes this way more approachable than whisper-based alternatives i've tried. one thing that would actually push me to use it daily though, could you add a custom vocabulary or replacement rule system? like so i can set "levitan" to always become "Levitan" or auto-format certain phrases i say a lot. would make the raw transcription feel way more polished without me having to fix stuff every time.
@kocanaoglu80506 I’m planning to add this together with a small local LLM that will handle punctuation, paragraph breaks, and dictionary-based word correction.
honestly love the offline angle, super rare these days. one thing i'd actually use though is a custom vocabulary or shortcuts feature so i can teach it my weird jargon and have it auto-replace spoken phrases with longer snippets, that would save me way more time than just raw dictation
@sendurgut14824 I’m planning to add this together with a small local LLM that will handle punctuation, paragraph breaks, and dictionary-based word correction.