Live:Local-first TTS · macOS

Select it.
Snipe it.
Hear it.

VoxSnipe turns any highlighted text into natural speech with one hotkey — and gives your AI tools a voice through a local webhook. No copy-paste into ten tabs. No account. Your words never leave your machine.

Private beta · macOS 13+ · be first in line, no spam.

  • 6 voice engines
  • Works offline
  • Text stays on your Mac
Voxsniper
Home
Voxsniper
Local-first TTS · running
Quick speak
Read aloud whatever is selected or in your clipboard
Speak from clipboard
SSpeak selectionXStop
Webhook
External apps can trigger speech via HTTP
127.0.0.1:7331POST /speakGET /health
Do Not Disturb
Speech is enabled
Mute speech…
Statistics
Lifetime activity
24
Speeches
2.6k
Characters
2.6
Minutes
Recent activity
Your latest speeches
The deploy finished — all 14 checks passed2m
Claude is done.1h
Markieren, kopieren, anhören.3h

Reads from anything you can select

Claude CodeChatGPTVS CodeSafariObsidianMailSlackNotionTerminalPDFsXcodeArc
Three ways to fire

One companion,
three triggers.

However you work, there's a way to start speech without breaking flow — a keystroke, an HTTP call, or a click in the menu bar.

Global hotkey

Highlight text in any app, press your hotkey, and VoxSnipe reads it aloud. Rebind it to anything in Settings.

speak selection

Local webhook

A tiny HTTP server on 127.0.0.1. Any tool — Claude Code, a script, a CI job — can make your Mac talk.

POST /speakGET /health:7331

Menu bar

Lives quietly in the tray. Speak the clipboard, pause, or mute for a while — all one click away, no window needed.

Speak clipboard
How it works

Reading costs time. Listening doesn’t.

From selected text to spoken words in under a second — no window to open, no flow to break.

  1. Highlight anything

    An error message, a paragraph, a paste from ChatGPT — select it wherever it lives.

  2. Press the hotkey

    VoxSnipe grabs the selection, cleans it up, and routes it to your chosen voice.

  3. Hear it instantly

    Natural speech, streaming as it generates. Stop, pause, or replay it from history.

Give your eyes a break. Listen to long threads, docs, and AI replies while you keep your hands on the keyboard.

Built for AI workflows

Give your tools
a voice.

VoxSnipe runs a local HTTP endpoint on your Mac. Point any tool at it and your machine speaks — no cloud relay, no exposed ports.

  • Have Claude Code say “done” the moment a task finishes
  • Let long builds, deploys, and test runs announce themselves
  • Send a different voice for AI output than for your hotkey
  • Trigger speech from any script, cron job, or CI pipeline
POST /speakGET /healthPOST /notify/<token>127.0.0.1:7331
~/.claude/settings.json
{
  "hooks": {
    "Stop": [{
      "hooks": [{
        "type": "command",
        "command": "curl -s -X POST http://127.0.0.1:7331/speak \
          -d '{"text":"Claude is done.", "source":"claude"}'"
      }]
    }]
  }
}
200 OK{ "status": "queued", "chars": 15 }
The voices

Six engines,
one surface.

Start free with the built-in macOS voices, then bring your own key for studio-grade cloud voices. Mix and match — VoxSnipe routes each source to the voice you pick.

Free

macOS built-in

System voices, fully offline.

Free · runs on this Mac

OpenAI

gpt-4o-mini-tts · 6 native voices · multilingual.

$0.06 / 1k chars

ElevenLabs

Premium quality, voice cloning, streaming.

from ~$5 / mo

Google Gemini

Native audio · 24 voices · 30+ languages.

Generous free tier · $0.05 / 1k

Azure Speech

500+ neural voices · 140 languages · custom voices.

500k chars free / month

Free

Custom endpoint

Bring your own OpenAI-compatible TTS server.

Self-hosted · your cost

Cloud voices are billed directly by the provider — VoxSnipe never marks them up or sees your usage.

Everything in the box

Small companion.
Serious control.

VoxSnipe stays out of your way until you need it — then gives you precise control over privacy, cost, and how things sound.

Private by design

The built-in voice runs entirely offline. API keys are stored in your local credential store and used only to call the provider you chose. No account, no analytics, no middleman — your text never passes through our servers, because there are none.

No accountNo telemetryKeys stay localOpen webhook on 127.0.0.1 only

Cost control

See a live dollar estimate per request and set budget limits per provider. Never a surprise bill.

ElevenLabs · this month$1.24 / $5.00

Voice routing

Send Claude one voice, your hotkey another, German and English their own — automatically.

Do Not Disturb

Mute speech for 5 minutes, an hour, or until you say so. Focus when you need it.

History & replay

Every spoken snippet is logged locally so you can replay it with a click — or clear it instantly.

Auto language

Detects the language and picks the right voice on the fly.

Skips code blocks

Reads prose, not punctuation soup. Code fences are quietly skipped.

Streams instantly

Audio starts as it generates — no waiting for the whole clip.

Pricing

Honest pricing. No subscription trap.

Start free with local voices. Unlock cloud voices and automation with a one-time Pro purchase — no monthly fee, ever.

Free

€0forever

Everything local, for as long as you like.

Join the waitlist
  • Built-in macOS voices
  • Global hotkey + menu bar
  • Local history & replay
  • Do Not Disturb
  • No account · fully offline
Most popular

Pro

€29one-time

Cloud voices and AI automation. Buy once, keep it.

Get early access
  • Everything in Free
  • All cloud engines — OpenAI, ElevenLabs, Azure, Gemini, custom
  • Local webhook for AI workflows
  • Voice routing per source & language
  • Cost control & budget limits
  • Updates for this major version

Team

Custom

Volume licensing for studios and companies.

Talk to us
  • Everything in Pro
  • Volume licensing & one invoice
  • Centralized billing
  • Priority support & onboarding

Cloud voices are billed separately by each provider · prices incl. VAT · Pro is a one-time purchase, not a subscription.

Good to know

Questions, answered.

With the built-in macOS voice, nothing leaves your machine at all. When you choose a cloud voice, only the text you ask for is sent straight to that provider using your own key — never through any VoxSnipe server, because we don’t run one.

Not for the free macOS voices. For OpenAI, ElevenLabs, Azure, Gemini, or a custom endpoint, you bring your own key and are billed directly by the provider. VoxSnipe stores keys in your local credential store and never sees your usage.

It depends on the engine — roughly $0.06 / 1k characters for OpenAI, Gemini has a generous free tier, and Azure includes 500k characters free each month. VoxSnipe shows a live estimate per request and lets you set a hard budget cap per provider.

VoxSnipe runs a small HTTP server on 127.0.0.1:7331. Add a Stop hook to your ~/.claude/settings.json that POSTs to /speak, and Claude Code will speak the moment a task finishes. Any script or CI job can do the same.

VoxSnipe targets macOS 13 Ventura and later, on both Apple Silicon and Intel. It’s a native desktop app — no browser, no Electron bloat.

Yes. You pay €29 once and keep Pro for good, including every update to this major version. No subscription, no auto-renewal. Cloud voice usage is the only ongoing cost, and that goes to the provider, not to us.

VoxSnipe is in private beta. Join the waitlist and we’ll email you the moment a build is ready for you — first come, first served.

Give your Mac a voice.

Be first to try VoxSnipe. Join the private beta and hear your first text the day it’s ready.

Private beta · macOS 13+ · no spam, unsubscribe anytime.