⚡ EYES-FREE AGENT HANDOFF

Stop Babysitting Your Agent's Terminal

Your agent builds in the background. VoiceFi speaks concise 5-second soundbites through your laptop speakers when it's done—then opens the mic so you can steer it hands-free.

Works with: Claude Code Google Antigravity Cursor OpenAI Codex Terminal CLI

Listening for AI Agents

Standing by for agent speech...

INTEGRATION TOOLKIT

⚡ Ways to Speak Aloud With VoiceFi

Pick a method to speak from terminal, agent, or browser:

⏰ The Acoustic Alarm Clock for Developer Tasks:

Each time task X finishes, say Y in Z voice. Zero install, zero dependencies.

railway up && curl -s "https://voicefi.org/api/tts?voice=viv&text=Deployment+is+live+on+Railway" -o /tmp/v.mp3 && afplay /tmp/v.mp3

🚀 Cloud & Host Deploys

Chain with && after railway up, vercel --prod, or wrangler deploy so you never have to watch dashboard spinning wheels.

⚡ Local Builds & Scripts

Chain after npm run build, docker compose, or add to package.json: "build": "next build && curl -s ... && afplay /tmp/v.mp3".

• macOS: saves to /tmp/v.mp3 and plays instantly via native afplay. • Linux: replace with mpv - or aplay

🤖 Instruct Your AI Coding Agent to Speak:

Give this rule to Claude Code, Antigravity, or Cursor so it speaks soundbites when finishing long tasks.

"When you finish running tests, builds, or refactoring, speak a concise 5-second summary out loud through my laptop speakers using: curl -s 'https://voicefi.org/api/tts?voice=viv&text=TASK_SUMMARY_HERE' -o /tmp/v.mp3 && afplay /tmp/v.mp3"

• Paste into chat, your system prompt, or save in AGENTS.md or CLAUDE.md.

🌐 WebMCP Autonomous Browsing Agent Prompt:

For OpenAI Operator, Claude Computer Use, browser-use, or Puppeteer/Playwright agents.

"Go inspect voicefi.org/agent and surprise me by speaking your findings out loud through my laptop speakers using WebMCP."

• WebMCP Tools: voicefi_speak_to_developer & voicefi_listen_to_developer • Headless JS hook: window.playVocalAudio(voice, text)

🍏 Full Native Desktop Layer (0ms Offline Speech + HUD):

Installs native CLI, Dynamic Island HUD, and auto-hooks for Claude Code & Antigravity.

curl -fsSL https://voicefi.org/install.sh | bash

• Run: vifi speak "Task done" or vifi rc for mobile companion • Supports Apple Ava (Premium) 0ms offline neural synthesis

Hear It In Action (1-Click Auditions):

WORKFLOW ACCELERATION

Why VoiceFi? Workflow Simplification & The Conversational Loop

AI agents spend minutes compiling, testing, and refactoring in silent background terminals. VoiceFi closes the feedback loop: concise 5-second spoken soundbites when tasks finish, zero window babysitting, and continuous hands-free steering.

🎙️

The Conversational Loop

Continuous hands-free turn taking. Background agent finishes → speaks a 5s soundbite → chimes → auto-opens mic. Steer your agent without touching the keyboard.

Agent Soundbite → Mic Handoff → Execution

🔄

Cross-Agent Handoffs

Seamless multi-agent delegation. Antigravity delegates to Claude Code, or Cursor triggers a test runner. VoiceFi coordinates spoken context and return routing.

vifi send --to claude • 0 Window Switching

📱

Remote Companion

Step away from your desk. Connect your mobile phone via local WebSocket or PWA companion (vifi rc) to listen and steer while pacing or making coffee.

Full-Duplex Audio Relay • Mobile PWA

Silent Terminal Babysitting vs. VoiceFi Conversational Loop

✕ Silent Agent Terminal

Constant alt-tabbing & window babysitting
Trapped reading walls of text to verify simple tasks
Hands tied to keyboard for every follow-up instruction
Stuck at desk waiting on builds, tests, and deploys

✓ VoiceFi Conversational Layer

Eyes-free awareness: acoustic notifications the instant tasks finish
5-second soundbite summaries (<20 words) with zero fluff
Synchronous auto-handoff: speak your next thought hands-free
Desk-free freedom: pace or step away with mobile companion

Autonomous Agent Console DOM & Vision

Voice:

🎁 Spoken Surprise Link: Auto-speaks on click

Agent:

Voice:

Message:

Share on X ↗

LIVE WEBMCP INVOCATION LOG 0 events

Waiting for an AI agent to call voicefi_speak_to_developer...

EXPERIENCE THE FULL ACOUSTIC LAYER

Take VoiceFi Beyond The Browser Tab

What you heard here is just the web surface. The complete desktop layer directs your AI agents with 0ms offline Apple Silicon speech, native Dynamic Island HUD, and full-duplex conversational barge-in.

Demo Product →

Open Source CLI 100% FREE / MIT

Native Python CLI tool, 0ms Apple Ava neural voice, Silero neural VAD, Stdio JSON-RPC 2.0 MCP server, and terminal hooks for Claude Code and Antigravity.

View GitHub Repository ↗

curl -fsSL https://voicefi.org/install.sh | bash

🍏 macOS App & Pro FREE CORE FOREVER + PRO TRIAL

Floating AppKit Dynamic Island HUD, menu bar tray, ⚡ Speed Talking (1.25x–3.0x), 1-click macOS installer (.dmg), mobile PWA companion sync, and multi-agent soundboard.

Get Mac App (.dmg) →