Echoic

ECHOIC

Remember Everything
Said

Meeting transcription and system-wide dictation.100% on-device. No cloud. No compromise.

Download for macOSSee how it works

macOS 14.2+ · Apple Silicon recommended

Live Transcription

Two tools. One app.

The only macOS app combining meeting transcription and system-wide dictation

Team Standup
REC 12:34
Me
Let's discuss the Q2 roadmap priorities...
Alex Chen
I think we should focus on the API first...
2 speakers · Live
Meeting Mode

Meeting Transcription

Capture every conversation with speaker-labeled transcripts, powered entirely on your device.

  • 🎙️

    Dual-Stream Capture

    Records microphone and system audio simultaneously — your voice and everyone else, cleanly separated.

  • 💬

    Live Transcript

    Watch the conversation unfold in real time with speaker-labeled chat bubbles.

  • 👥

    Speaker Diarization

    Identifies and separates speakers into "Me" and "Them" channels with inline renaming.

  • 🧠

    AI Summaries & Action Items

    Get meeting summaries and extracted action items from any AI provider you choose.

  • 📅

    Calendar Integration

    See upcoming meetings in the menu bar. Quick-start recording from any calendar event.

  • 📤

    Multi-Format Export

    Export as Markdown, DOCX, or PDF. Markdown is Obsidian-vault compatible.

  • 💬

    Per-Meeting AI Chat

    Coming Soon

    Ask follow-up questions about any transcript — "what were the action items?" or "summarize in Spanish".

  • 🔍

    Cross-Meeting Search

    Coming Soon

    Ask questions across all your past meetings. Fully offline via local embeddings.

Dictation Mode

System-Wide Dictation

Speak into any app on your Mac with sub-500ms latency — no cloud, no delay.

  • ⌨️

    Type Into Any App

    Hold your hotkey, speak, and text appears at your cursor — in VS Code, Slack, Terminal, anywhere.

  • ⚡

    Sub-500ms Latency

    Near-instant transcription with local neural models. No round-trip to the cloud.

  • 🎯

    Voice Commands

    "New line", "delete that" — natural editing commands built in.

  • 🔄

    Concurrent Mode

    Dictate while recording a meeting — both modes run simultaneously without interference.

  • 🎛️

    Push-to-Talk or Toggle

    Hold to speak, or toggle for hands-free dictation. Your preference.

  • 📋

    Dictation History

    Searchable history of everything you've dictated, always at hand.

  • ⚙️

    Custom Voice Commands

    Coming Soon

    Define your own commands — "insert signature" expands to any configured text block.

  • 🎨

    Per-App Profiles

    Coming Soon

    Different dictation behavior per app — formal prose in email, code syntax in VS Code.

Dictating...
...scheduled for

Hold ⌥ to dictate into any app

Privacy First

100% On-Device.Zero Cloud.

Every word stays on your Mac. Your meetings, your voice, your data — never transmitted, never stored on a server.

No Audio Uploaded

Your voice never leaves your Mac. All speech recognition runs on your device using CoreML.

Compliance-Friendly

On-device processing makes Echoic safe for sensitive conversations — legal, medical, financial.

All speech processing uses CoreML · Runs entirely on your Apple Silicon chip
BYO AI Provider

Your AI, Your Rules

Route different features to different providers. Use Ollama for full privacy, or cloud models for maximum quality. You're in control.

Ollama logo
Local

Ollama

Fully local, 100% offline AI

OpenAI logo
Cloud

OpenAI

GPT-4o and beyond

Anthropic logo
Cloud

Anthropic

Claude 3.5 Sonnet

Groq logo
Cloud

Groq

Ultra-fast inference

xAI logo
Cloud

xAI

Grok models

Apple Intelligence logo
Local

Apple Intelligence

On-device Apple AI

What AI can do in Echoic:

Meeting SummariesAction Item ExtractionCustom AI TemplatesDictation CleanupPer-Feature RoutingStreaming Generation

Fully offline mode available

No internet connection required

Use Ollama locally — zero data leaves your Mac

Get Started in Minutes

How It Works

From download to dictating in under 3 minutes.

1

Download & Install

Download the DMG, drag Echoic to Applications. It lives in your menu bar.

2

Grant Permissions

Allow Microphone and Accessibility access. A friendly guided setup walks you through it.

3

Start Talking

Hold your hotkey to dictate, or click Record to capture any meeting. That's it.

What you'll need

  • ✓A Mac running macOS 14.2 or later
  • ✓Apple Silicon recommended (Intel supported)
  • ✓~600 MB free storage for speech model
  • ✓Microphone + Accessibility permissions (guided setup)

Speech Engine

Choose Your Engine

Every engine runs locally on your Mac. Pick based on your language, speed, and quality needs.

⭐ Recommended

Parakeet v3

Best
⚡Fast💾~542 MB🌐English

NVIDIA's best-in-class English model, optimized for Apple Silicon

Whisper Large

Great
⚡Fast💾~632 MB🌐90+ languages

OpenAI's multilingual model — transcribes any language

Moonshine v2

Good
⚡Fastest💾~290 MB🌐English

Smallest and fastest — perfect for real-time low-latency use

Apple SpeechAnalyzer

Good
⚡Fast💾0 MB (system)🌐System language

Built into macOS — zero download, powered by Apple's framework

All engines download once and run offline. Switch anytime in Settings → Voice Engines.