Your voice. Your screen. One AI.
Talk to AI agents instead of typing — explain more context in seconds than you'd ever write. Voice-type polished text in any app, or get instant screen-aware help. No screenshots. No copy-pasting. Just speak.
Same thought. Different speed.
You think at 150 words per minute. Your keyboard caps you at 40. When you type to AI, you self-censor — you skip context because it's slow to write. With voice, you express the full picture in seconds.
Watch both prompts fill in real time — same 15 seconds, different results.
3x more expressed. Same time.
8 steps. Or one.
- See bug on screen
- Screenshot it
- Open ChatGPT
- Paste screenshot
- Type context
- Wait for response
- Read the answer
- Switch back to your app
"What's wrong here?"
↓Answer in seconds. Right where you are.
Up and running in 60 seconds.
Download & install
Drag to Applications. Grant mic + accessibility permissions.
Press your hotkey. Speak. Release.
A floating island shows you it's listening. Speak naturally.
Polished text appears
Wherever your cursor is. That's it. No copy-paste. No switching apps.
Three modes. One hotkey.
Voice Typing
Speak → polished text appears
AI Command
Ask anything → instant answer
AI Workspace
Deep work → persistent AI chat
Write 3x faster. Get help instantly.
Write faster
Speak your thoughts. Get polished text.
- 3x faster than typing — 150 WPM speaking vs 40 WPM typing
- AI adapts tone per app — professional in Mail, casual in Slack
- Express your full thought without self-editing mid-sentence
- Output is better than what you'd have typed — richer input, smarter AI
Get instant help
It sees your screen. Just ask.
- No more screenshots → paste into ChatGPT → wait → read
- Stuck on an error? "What's wrong here?" — it already sees it
- Lost in a complex UI? "Guide me" — step-by-step, in plain language
- Works in any app: code editors, browsers, settings, forms
Every AI tool is blind. Auris sees your screen — only when you ask.
Other tools make you screenshot, paste, and explain. With Auris, just ask. Screen capture is opt-in, never automatic, and nothing is stored — your screenshot is processed and immediately discarded.
Thread 1: Fatal error: Index out of range Professional in Mail. Casual in Slack. Raw in Terminal.
Same voice input, different output per app. Fully customizable — set each app to your preferred style.
Give it 3 days.
Every Auris user says the same thing: "I didn't think I needed this." Then after a few days of speaking instead of typing, going back feels like switching from autocomplete to a typewriter.
Speaking removes the friction that makes you write shorter, simpler messages. Your communication gets richer.
Average typing: 40 WPM. Average speaking: 150 WPM. That's 3x faster — backed by research.
Your spoken-then-refined text reads better than what you'd have carefully typed — because you gave the AI more to work with.
The more you use it, the more Auris learns your patterns. Per-app settings dial in. It gets better every day.
Private by default. Free to start.
Most voice apps stream your audio to the cloud for transcription — your voice, your biometric identity, on someone else's server. Auris transcribes entirely on your Mac. Your audio never touches the internet. Not even for a millisecond.
On-device Core ML model runs locally. 25 languages. No audio ever transmitted.
AI refinement sends only text — never audio. Processed and discarded immediately. Nothing logged or stored in the cloud.
Only captured when you explicitly ask. Never automatic. Processed and discarded — no screenshots stored anywhere.
Frequently Asked Questions
Is Auris better than macOS built-in dictation?
Yes. Apple's dictation is basic transcription only. Auris adds AI text refinement (fixing grammar, tone, and formatting), per-app intelligence, screen-aware AI help, and voice-to-AI-agent workflows — all while keeping transcription on-device and private.
Can I use Auris to talk to AI agents like Claude, ChatGPT, or Cursor?
Absolutely. Auris works in any text field on macOS. Speak naturally to Claude Code, ChatGPT, Cursor, Copilot, or any AI tool — explain more context in seconds than you'd ever type, and get better results on the first try.
Is my voice data private? Does Auris send audio to the cloud?
Voice transcription is 100% on-device using Apple's speech framework — no audio ever leaves your Mac. AI refinement (optional) sends only the text to process, never the original recording. Screen capture is opt-in and immediately discarded after processing.
What makes Auris different from other Mac dictation apps?
Auris combines three things no other app does: (1) AI-powered text refinement that adapts tone per-app, (2) screen-aware AI that can see and help with anything on your display, and (3) voice-first AI agent workflows for talking to Claude, ChatGPT, and coding tools.
Does Auris work offline?
Voice transcription works fully offline — it runs on-device with no internet needed. AI refinement and screen-aware features require an internet connection for AI processing, but your audio never leaves your Mac.
Which macOS versions does Auris support?
Auris requires macOS 14 (Sonoma) or later and Apple Silicon (M1/M2/M3/M4). Download is free with a 7-day Pro trial — no credit card required.