Skip to content

AnsCodeLab/VoxMemo

BranchesTags

Repository files navigation

VoxMemo

Local-first meeting recorder — record, transcribe, identify speakers, and summarize meetings entirely on your machine. No cloud. No subscriptions.

Built with .NET 10 and Avalonia UI. Runs on Windows, Linux, and macOS.

Screenshots

Recording view
Recording — Live captions while you record 		Meetings view
Meetings — Transcript, speakers, and AI summary
Smart Process pipeline
Smart Process — One click: transcribe → identify speakers → summarize 		Settings view
Settings — Ollama, OpenAI, Anthropic, or any OpenAI-compatible API

Why VoxMemo?

VoxMemo Otter.ai / Fireflies Local Whisper scripts
Privacy 100% local Uploads to cloud Local
Cost Free $10–$20/mo Free
Speaker ID ✅ AI-powered ✅ ❌ manual
Custom AI models ✅ Ollama, LM Studio ❌ ❌
Works offline ✅ ❌ ✅
Desktop UI ✅ Web/mobile ❌ CLI
Live captions ✅ ✅ ❌

Features

Recording

  • Mic + Speaker mix — records both your voice and remote participants in one file (ideal for Zoom/Teams/Meet)
  • Record from microphone, system audio, or both simultaneously
  • Live captions while recording (Whisper tiny model, auto-downloaded) — new line per sentence, selectable and copyable
  • Pause/resume recording
  • Global hotkey (default Ctrl+Shift+R) to start/stop from anywhere
  • System tray with full controls: start, stop, pause, audio source, device, language

Transcription

  • Local speech-to-text powered by Whisper.net — no cloud required, no data sent anywhere
  • Auto language detection — Whisper automatically detects language, including mixed Vietnamese/English conversations
  • Supports 30+ languages including English, Vietnamese, Chinese, Japanese, Korean, French, German, Spanish
  • Import and use custom Whisper models (any ggml-format .bin file)
  • Auto-converts imported audio (MP3, M4A, OGG, FLAC, etc.) to Whisper-compatible WAV

AI Processing

  • Smart Process — one-click pipeline: Transcribe → Identify Speakers → Summarize
  • Speaker identification — AI reformats transcript as dialog with named speaker labels, shown in a dedicated Speakers tab (original transcript preserved)
  • Meeting summarization — structured summaries referencing speakers by name
  • Auto-title generation — AI creates descriptive meeting titles from content
  • Custom AI prompts — override summary and speaker identification instructions
  • Prompt templates — action items, executive briefs, technical reviews, interviews, standups, brainstorms
  • Supports Ollama (local), OpenAI, Anthropic, LM Studio, and any OpenAI-compatible API

Meetings Management

  • Import audio — upload existing MP3, WAV, M4A, OGG, FLAC, WMA, AAC files
  • Import transcript — paste text or load a .txt file to skip recording entirely
  • Audio player with play, pause, stop, and seek slider — state preserved when switching tabs
  • Editable transcripts — correct text before summarizing
  • Copy and export — clipboard or file export for transcript/summary/audio
  • Search meetings by title, platform, or date
  • Language badge — per-meeting language tag, editable after recording

Job Queue

  • Background processing queue — all heavy work off the UI thread
  • Real-time progress — status, provider/model info, elapsed timer per job
  • Cancel active jobs or remove queued ones
  • Toast notifications on completion/failure
  • Jobs process sequentially in guaranteed order

First-Run Setup

  • Onboarding wizard on first launch — configure AI provider, download a Whisper model, and select audio device in one guided flow
  • Re-run anytime from Settings → Re-run Setup Wizard

Settings

  • AI provider configuration with live model refresh
  • Whisper model management — download built-in models, import custom .bin files
  • GPU acceleration — Vulkan support for Intel Arc and AMD GPUs
  • AI timeout — configurable 1–60 minute timeout per request
  • Automation — auto-transcribe and auto-summarize after recording
  • Configurable languages — add/remove from 30+ Whisper-supported languages
  • Global hotkey — customizable (Windows only)
  • Start with Windows — auto-launch via registry
  • Storage location — choose where recordings are saved

Download

👉 Latest release — v1.5.0

Platform Download
Windows (Installer) VoxMemo-v1.5.0-Setup.exe
Windows (Portable) VoxMemo-v1.5.0-win-x64.zip
Linux x64 VoxMemo-v1.5.0-linux-x64.tar.gz
macOS Intel VoxMemo-v1.5.0-osx-x64.tar.gz
macOS Apple Silicon VoxMemo-v1.5.0-osx-arm64.tar.gz

Prerequisites

Requirement Windows Linux macOS
ffmpeg + ffprobe Optional (import only) Required Required
PulseAudio (parecord) — Required (recording) —
ffplay — Recommended (playback) —
notify-send — Recommended (notifications) —
BlackHole — — Required for system audio
Ollama Optional Optional Optional

Install dependencies

Linux (Debian/Ubuntu):

sudo apt install ffmpeg pulseaudio-utils libnotify-bin

Linux (Fedora):

sudo dnf install ffmpeg pulseaudio-utils libnotify

Linux (Arch):

sudo pacman -S ffmpeg pulseaudio libnotify

macOS:

brew install ffmpeg
# For system audio capture, install BlackHole:
# https://github.com/ExistentialAudio/BlackHole

Getting Started

1. Download a Whisper model

Go to Settings → Transcription and download a model:

Model Size Best for
tiny ~75 MB Live captions (auto-downloaded)
base ~150 MB Fast transcription
small ~500 MB Good balance
medium ~1.5 GB High accuracy
large-v3-turbo ~1.5 GB Best for Vietnamese / Asian languages
large-v3 ~3 GB Maximum accuracy

2. Set up AI (optional, for summarization)

Ollama (local, free — recommended):

ollama pull gemma3        # fast, good quality
ollama pull qwen3:8b      # great for Vietnamese

Ollama runs at http://localhost:11434 by default.

LM Studio / LocalAI / vLLM: Select "OpenAI" provider, set Base URL (e.g. http://localhost:1234/v1), leave key empty.

OpenAI / Anthropic: Enter your API key in Settings → AI Provider.

3. Record a meeting

  1. Select "Both (Mic + Speaker)" as audio source for full meeting capture
  2. Click Start Recording (or press Ctrl+Shift+R)
  3. After the meeting, click Stop
  4. Click Smart Process to transcribe, identify speakers, and summarize in one step

Build from Source

git clone https://github.com/AnsCodeLab/VoxMemo.git
cd VoxMemo
dotnet restore
dotnet run

Run tests:

cd VoxMemoTests
dotnet test

Build all platforms:

./build.sh

Tech Stack

Component Library
UI framework Avalonia UI 11.3
MVVM CommunityToolkit.Mvvm 8.2
Audio (Windows) NAudio 2.2 — WASAPI + WaveOut
Audio (Linux) PulseAudio + ffmpeg
Audio (macOS) ffmpeg avfoundation + afplay
Transcription Whisper.net 1.8 (local, all platforms)
GPU acceleration Whisper.net.Runtime.Vulkan
Database EF Core + SQLite 9.0
Logging Serilog 4.x
Runtime .NET 10

Supported Platforms

Platform Recording System Audio Playback Hotkey Auto-start
Windows 10/11 x64 NAudio WASAPI Loopback capture NAudio ✅ Ctrl+Shift+R Registry
Linux x64 PulseAudio Monitor source ffplay — XDG autostart
macOS x64/arm64 ffmpeg BlackHole required afplay — LaunchAgent

Logs

Platform Path
Windows %AppData%\VoxMemo\logs\
Linux ~/.config/VoxMemo/logs/
macOS ~/Library/Application Support/VoxMemo/logs/

License

MIT License — free to use, modify, and distribute.

Author

AnsCodeLab — github.com/AnsCodeLab

About

No description, website, or topics provided.

Resources

Readme

Contributing

Contributing
Activity

Stars

3 stars

Watchers

0 watching

Forks

0 forks
Report repository

Releases 6

v1.5.0 Latest
Apr 30, 2026
+ 5 releases

Packages

 
 
 

Contributors

Languages