NewVoice input app

Speak clearly. Adapt to context.

Local-first voice input for Mac—cleanup, rewrite, translation, and per-app style from one tap-to-talk habit. Voxt is more than speech-to-text.

EMLNJC

Mac users and counting

Voxt desktop home preview
VoxtFN
14:12100%
Mac only · Local-first

Don’t type, just speak

Tap a shortcut, talk naturally, and Voxt turns speech into clear writing in every app

Voice notes
⌘ + Space
  • Draft launch checklistTodo
  • Capture follow-ups from standupDoing
  • Paste notes into LinearDone

VOICE NOTE CAPTURE

Wake Voxt anywhere, capture to-dos in seconds, then paste or drag them into your apps—task management without the friction.

Query All User
SQL Transform
AI
Natural language → SQL
Action · SQL
SELECT *
FROM users;

DIFFERENT APPS, DIFFERENT RESULTS

Configure AI transforms for any app or URL—spoken text becomes the right result instantly, and your throughput doubles.

你好
Hello
こんにちは
안녕하세요
Bonjour
Hola
Hallo

LANGUAGE ON THE FLY

100+ languages on tap. Translate wherever you are—read and communicate without barriers.

Zoom · Product sync
34:12
YYou
AAlex
MMia
JJon

Speakers

  • Alex02:14

    Let’s lock the timeline.

  • Mia02:31

    I’ll share the notes.

  • Jon03:02

    Summary looks good.

MEETING NOTES & SUMMARIES

Speaker timestamps, live translation, live extraction, and summaries you can reuse anytime.

Your voice powers every workflow

Cleanup, rewrite, translation, chat, notes, per-app enhancement, live meetings, and recording analysis—eight flows in one menu bar app.

Voxt cleans what you say

Tap fn and talk, tap again to finish. Voxt drops filler, resolves self-corrections, fixes punctuation and numbers, and pastes at your cursor.

  • Live transcript in the overlay while you speak
  • Filler words and false starts get filtered out
  • “Friday—no, Monday” lands as just “Monday”
  • Switch to Clear Structure when you want lists

Voxt rewrites the meaning

Select text, tap fn+Control, and say how to change it. The result opens in an answer card—inject it to replace the selection.

  • Say “make it shorter” or “more direct” in natural language
  • Rewrite only the selection when you need surgical edits
  • Preview the result before it replaces the draft
  • Inject, copy, or continue refining from the card

Voxt translates what you select

Select text and tap fn+Shift—no speaking needed. The translation opens in a result card, ready to replace the selection.

  • Translate selected text without opening another app
  • Switch the target language right on the result card
  • Or speak, and the translation is pasted directly
  • Its own model, prompt, and dictionary terms

Voxt opens an AI chat

With nothing selected, what you say is the prompt. The answer opens in a card—press Continue to keep the conversation going by voice.

  • No selection? Your speech becomes the prompt
  • Continue (or Space) adds a follow-up in the same thread
  • Inject into the current input or copy the answer
  • Only what you say or select is sent—never your screen

Voxt captures your notes

Tap right ⌘ to open the floating notes panel, tap again and speak. Voxt saves the note with a short AI title, ready to move from To do to Done.

  • Peek at notes by resting the pointer in a screen corner
  • Every note gets a short AI-written title
  • To do, In Progress, and Done—with priorities
  • Sync to Obsidian or Reminders, or drag into any app

Voxt enhances per app

App Branch swaps the transcription prompt by app or URL. Same fn tap—chat stays casual, email gets a subject and sign-off, the terminal gets real flags.

  • Group apps or URL patterns, each with its own prompt
  • Start from templates: Chat, Email, IDE / Terminal, Docs
  • Spoken “dash m” becomes -m in the terminal
  • The overlay shows the matched app while it works

Voxt transcribes your meetings

Tap fn+Option and a live card follows the call—Me and Them on one timeline. Finish to save it and open Meeting Details with a summary and to-dos.

  • Your mic is Me, system audio is Them—live, side by side
  • Optional live translation for what Them says
  • Meeting, Subtitles, or Recording capture modes
  • Meeting Details: playback, summary, TODOs, follow-ups

Voxt analyzes your recordings

Drop an audio or video file into Files. Voxt prepares the audio, transcribes it, identifies speakers, and saves a full meeting analysis.

  • MP3, MP4, M4A, WAV, MOV, and more—up to 12 hours
  • A task queue with progress, Prioritize, Cancel, and Retry
  • Speaker separation labels Speaker 1, Speaker 2…
  • Runs on local MLX Audio or your own remote ASR

Built for every workflow

Coding

Ship notes without breaking flow

Dictate commit messages, PR summaries, and inline comments—then paste straight into your editor with technical terms intact.

Writing

Draft faster than you type

Speak long-form ideas, then rewrite for tone and clarity. Email, docs, and blogs share one voice entry point.

Meetings

Never lose what was said

Capture decisions, action items, and speaker notes in real time—then translate or summarize when the call ends.

Messaging

Reply in the right register

Slack stays short, email stays polished, and chat stays natural—App Branch switches style without changing habits.

Multilingual input coverage

Speak any language

Voxt is built for Chinese–English mixed speech, cross-language translation, and multi-region teams. Transcription, translation, rewrite, and app enhancement share one set of shortcuts.

France
Germany
USA
China
Mexico
Japan
Brazil
Canada
UK
Colombia
UAE
95+ More
Plans & Pricing

Free local forever.

Download Voxt free, keep processing local, bring your own keys for remote models, and optionally support development with a one-time $8 thanks.

Ethan LuoMina ParkLeo ShenNadia RossJonas Wei

4.9/5 from early Voxt users

Local forever free — optional one-time support

Free plan

Perfect for individuals who want polished voice input across every app.

$0forever / local-first
  • Unlimited local transcription
  • Cleanup, rewrite, and translation flows
  • App Branch per-app enhancement rules
  • Personal dictionary & mixed-language
  • Bring your own remote API keys

Support plan

Optional one-time support for the indie developer behind Voxt.

$8one-time payment
  • Everything in Free forever
  • One-time thanks — not a subscription
  • Help fund ongoing Mac app development
  • Priority consideration for feature requests

Quick start

Up and running in 5 minutes

1Download Voxt
2Connect a model
3Start outputting
Yes—completely free.

Voxt is a macOS voice input app that turns speech into text you can act on. Tap a shortcut, talk, and Voxt cleans filler, finishes punctuation, then pastes—or continues into translation, rewrite, AI chat, and per-app enhancement without leaving your current field.

Ethan Luo

Ethan Luo

@ethanluo · Apr 24

App Branch finally clicked. Slack short, email polished, Linear structured—same tap-to-talk 🔥 zero app switching.

193421418.4K
Nadia Ross

Nadia Ross

@nadiaross · Apr 17

Select → translate ✨ keep writing. Mixed CN/EN notes land clean without opening another tab.

9161419.6K
Yuki Tan

Yuki Tan

@yukitan · Apr 10

Voice → AI chat ⚡️ paste into Cursor. Drafted three PR summaries this morning without leaving the editor.

162518413.2K
Mina Park

Mina Park

@minapark · Apr 22

I ramble. Voxt cleans. Done.

121816712.1K
Jonas Wei

Jonas Wei

@jonaswei · Apr 15

Local when privacy matters 🔒 remote when I want a sharper rewrite. Same shortcut. Feels like a Mac tool, not a browser demo.

142217611.9K
Marcus Chen

Marcus Chen

@marcuschen · Apr 8

Dictating into Notion used to dump a wall of text. Now Voxt drops bullets and action items at the cursor—meeting notes actually survive the call, and I stop rewriting the same mess after hangup.

182920616.4K
Leo Shen

Leo Shen

@leoshen · Apr 20

ASR and LLM as separate channels is such a quiet flex. Whisper (local) for capture, then Claude or Gemini only when I want a stronger rewrite. My keys, my routing—feels like a real desktop stack, not another SaaS upsell.

212719315.8K
Clara Boone

Clara Boone

@claraboone · Apr 13

“Make it shorter.” That’s it. Stays in the field I’m already in—no ChatGPT tab roulette.

112015810.7K
Iris Morgan

Iris Morgan

@irismorgan · Apr 6

Free + local-first 🙌 BYO keys when I need remote. No Voxt bill.

132117312.8K

Let thought sing, hands-free

Voxt turns your voice into polished text in seconds. No manual editing needed—rambling thoughts become clear, structured copy with perfect punctuation, minus the filler and typos.