Debrief

[dih-breef] noun • macOS app

A meeting notes app that runs entirely on your machine. Transcribes, summarizes, and extracts action items. No cloud, no servers. Just files you own forever.

from envie studio

Download for Mac
Weekly Standup - 23m

Team discussed Q1 roadmap priorities. Design review scheduled for Thursday. API documentation needs updating before launch.

Send updated wireframes to Mike
Schedule design review - Thu 2pm
Review API documentation

Works with everything

Zoom
Google Meet
Slack
MS Teams

Records any audio source on your Mac - browser tabs, desktop apps, or microphone input.

Recording

Capture every word.

Record system audio, mic, or both simultaneously. Works with Zoom, Google Meet, Slack huddles, MS Teams - anything that makes sound on your Mac.

Lives quietly in your menu bar. One click to start, one click to stop. The audio file is yours to keep forever.

Recording

00:23:45

Recording
Transcript

Sarah

"Let's discuss the Q1 roadmap priorities and what we need to ship."

Mike

"I think we should focus on the API first..."

Transcription

Every word, every voice.

Powered by OpenAI Whisper and NVIDIA Parakeet models running locally. Speaker identification tells you who said what. No internet required.

Intelligence

Skip the replay.

Local LLMs like Qwen and Llama extract what matters: key decisions, action items, and who's responsible. Never miss a follow-up again.

Everything saved as clean Markdown files you can open anywhere - Obsidian, Notion, VS Code, or just TextEdit.

Weekly Standup

Summary

Team aligned on Q1 priorities. Design review Thursday. API docs needed before launch.

Action Items

Send wireframes to Mike
Schedule design review - Thu 2pm
Review API documentation
Download for Mac

Your choice of AI models.

Pick the model that fits your needs. All run locally on your Mac - no cloud, no API keys, no per-minute fees.

Transcription Models

Speech-to-text powered by OpenAI Whisper

Tiny

Size~40 MB
SpeedFastest
AccuracyGood

Quick notes, drafts

Base

Size~150 MB
SpeedFast
AccuracyBetter

Everyday use

Small

Size~500 MB
SpeedFast
AccuracyGreat

Important meetings

Large v3

Size~3 GB
SpeedSlow
AccuracyExcellent

Maximum accuracy

Summarization Models

Local LLMs for summaries and action items

Llama 3.2 1B

Size~700 MB
SpeedFastest
QualityBasic

Quick summaries

Qwen3 4B

Size~2.5 GB
SpeedFast
QualityGreat

Best speed/quality balance

Qwen3 8B

Size~4.5 GB
SpeedMedium
QualityExcellent

Great reasoning

Gemma 3 12B

Size~7 GB
SpeedSlow
QualityBest

16GB+ Macs

100% local. Forever yours.

Unlike Otter, Fireflies, and other cloud transcription services, Debrief runs entirely on your device. Your conversations never leave your machine. No servers, no subscriptions, no monthly fees.

Confidential meetings stay confidential. Legal calls stay privileged. Sensitive discussions stay private. Your data is yours alone.

No cloud Works offline Files you own

01

Menu Bar

Lives quietly in your menu bar. One click to record, always ready when you need it.

02

Record

Capture system audio, microphone, or both at once. Works with any app.

03

Transcribe

Whisper and Parakeet models turn speech to text locally on your Mac.

04

Identify

Speaker diarization knows who said what throughout the conversation.

05

Summarize

Local LLMs extract key decisions, topics discussed, and what matters.

06

Action Items

Automatically pulls out tasks, deadlines, and who's responsible.

07

Markdown

Clean .md files you can open in Obsidian, Notion, or any text editor.

08

Offline

No internet required. Works on planes, in bunkers, wherever you are.

09

Own

Your files, your device, your data. No subscriptions, no cloud, forever yours.

One-time purchase

Own it forever.

No subscriptions. No cloud fees. No recurring charges.
One payment, yours for life.

$49 $19 USD

One-time payment

3 Macs Lifetime updates
Download for Mac

Everything included

  • Lifetime license - pay once, own forever
  • Use on up to 3 Macs
  • All future updates included
  • Local transcription models
  • Local summarization models
  • Speaker identification
  • Action item extraction
  • Markdown files you own