Private AI meeting notes for iPhone and Android

Record the meeting. Keep the intelligence. Lose the cloud risk.

Debrief captures audio, transcribes speech, and turns the conversation into summaries, action items, key decisions, and topic tags directly on the device. Nothing needs to be uploaded to a third-party server.

  • 100% on-device processing
  • Live transcription in 30-second chunks
  • Structured notes after every meeting
  • Free on App Store and Google Play
0

required cloud uploads

30s

chunk-based live transcription cadence

60m

meeting length supported in the current app flow

5

Whisper model sizes for different devices

Product story

Built for teams that cannot afford casual data leakage.

Debrief is not another cloud meeting tool built around uploads and policy-page caveats. It is designed for founders, operators, consultants, and privacy-sensitive organizations that want strong notes without handing raw audio to someone else's infrastructure.

01

Local recording and transcription

Audio capture, speech-to-text, and summary generation run on the device using local AI models.

02

Live transcript while the meeting is happening

Debrief processes recording chunks as they arrive so users can monitor the conversation instead of waiting until the end.

03

Notes people can actually use

Output includes a title, summary bullets, action items, key decisions, topic tags, and full transcript history.

04

Model control stays in the user's hands

Users can choose the Whisper model that fits their device and add a local LLM when they want deeper summaries.

05

Search, review, export, share

Meetings remain searchable on the device and can be exported in plain text or Markdown without changing the privacy model.

06

Works where connectivity is weak

Once the models are downloaded, the core workflow is designed around local execution instead of a permanent internet dependency.

How it works

A clear three-step flow from capture to decision.

The workflow is simple: install the right models for the device, record locally, and process the meeting into structured notes when the conversation ends.

1

Install the models once

Download a Whisper transcription model for the device. Add the local summary model to unlock action items, decisions, and richer notes.

2

Record with live status and transcript updates

Start recording in seconds, capture participant context, and watch transcript chunks update as the meeting progresses.

3

Review summaries and ship the follow-up

Debrief turns the transcript into a concise title, bullet summary, action items, key decisions, and searchable meeting history.

Device Fit

On-device AI is more private, but it is not identical on every phone.

Debrief is intentionally transparent about the tradeoff. Because transcription and summaries run locally, results depend on RAM, free storage, and how large a model the device can comfortably handle. The app recommends a transcription model based on device memory, but newer phones will deliver the strongest overall experience.

Best experience

Recent iPhones and flagship Androids

8 GB RAM class or better

  • Best fit for Whisper Medium or Whisper Large Turbo
  • Best chance of handling the optional 2.5 GB summary model well
  • Strongest live transcript responsiveness and richest note generation

Strong everyday use

Modern phones with solid memory headroom

6 GB RAM class

  • Good fit for Whisper Medium in many everyday workflows
  • Summary model may still be practical if storage headroom is available
  • Good balance for regular meeting capture and structured notes

Lighter local processing

Older phones and budget devices

2 GB to 4 GB RAM class

  • Start with Whisper Base or Whisper Small
  • Expect slower transcription and less room for large models
  • Best for recording-first workflows and smaller local setups

Fallback path

If you skip the larger models at first

Still usable, with reduced AI depth

  • Recording still works without a Whisper model installed
  • Without Whisper, live transcription is disabled
  • Without the summary model, Debrief falls back to a simpler summary path
75 MB-950 MB

Whisper model download range

2.5 GB

Optional local summary model

~30s

Live transcript chunk cadence in the current flow

RAM-based

In-app model recommendation logic

Transparency matters here.

Debrief is best used on newer phones with enough free storage and memory for local models. That is the cost of keeping audio and AI processing on the device instead of shipping private meetings to a server. The privacy gets stronger, but the hardware matters more.

Why Debrief wins

Privacy-first by architecture, not by a promise buried in fine print.

Debrief keeps its strongest claim simple. The important work happens on the device. That gives buyers a clearer answer to privacy, ownership, and compliance questions than a standard upload-first workflow.

Typical cloud tools Debrief
Raw meeting audio is uploaded for processing. Audio stays local unless the user explicitly exports it.
Product value depends on another company's infrastructure. Core value comes from local inference and local storage.
Privacy reassurance is usually policy-based. Privacy reassurance is built into the workflow itself.
Users get transcript-heavy output. Users get summaries, action items, decisions, and tags.

Best-fit buyers

Position Debrief for the people who care about memory and confidentiality.

Founders and product leads

Capture customer calls, roadmap reviews, and standups without adding another cloud system to the stack.

Consultants and client services teams

Keep sensitive conversations close to the device while still walking away with usable notes and follow-up actions.

Operations and field teams

Work in environments where connectivity is inconsistent and local-first tooling is more reliable than server-dependent workflows.

Privacy-sensitive organizations

Use Debrief as the simple answer when leadership asks where the audio goes and who can access it.

Download

Install Debrief directly from the store.

If Debrief is live in the App Store and Play Store, normal users should download it directly there. Email should be reserved for business use, partnerships, privacy questions, team rollout discussions, or product feedback.

App Store

For iPhone users

  • Best path for individual users on iPhone
  • Send normal user traffic straight to install
  • Removes unnecessary access-request emails
  • Replace the placeholder link with your real App Store URL
App Store

Business / Feedback

For higher-intent conversations

  • Use this for business inquiries, partnerships, or pilots
  • Use this for privacy requirements or team rollout questions
  • Use this for product feedback from serious users
  • Do not use this as the default path for normal app installs
Business / Feedback

FAQ

Answer the objections before the buyer asks them.

No. The product is positioned around local recording, local transcription, and local summary generation. Audio stays on the device unless the user explicitly exports content.

The app can generate a meeting title, summary bullets, action items, key decisions, topic tags, and a full transcript, all based on the recorded meeting.

Once the required models are downloaded, the core recording and processing flow is designed around on-device execution rather than constant internet access.

No. Debrief runs AI models on the device, so performance depends on RAM, free storage, and the model size the phone can support. Newer devices will generally handle larger models and richer summaries better than older or budget phones.

The product codebase is a mobile app targeting iOS and Android, with the website acting as the marketing and sales layer for that product.

Launch CTA

Need more than a download link?

If someone wants to talk about business use, partnerships, team rollout, privacy requirements, or product feedback, email is the right CTA. For regular users, the website should push directly to the App Store or Google Play.