MEETINGIQ / FEATURED BUILD

Better questions. Better meetings.

The useful question often arrives after the meeting. MeetingIQ helps surface it while the conversation is still happening.

A desktop companion that listens to selected meeting audio, follows the discussion, and suggests questions worth asking.

Desktop prototypeWindows · Desktop Chrome

The browser examples are scripted. The real app runs separately on your computer.

MEETINGIQFICTIONAL EXAMPLE

A claim worth questioning.

SOMEONE SAYS

We can launch on Friday. I assume the security review is already done.

QUESTION RADAR / SAMPLE

Has the security review been completed and formally approved before committing to Friday?

Choose a statement to see a different example. No recording or AI calls.

BEHIND THE BUILD / PERSONAL PROJECT

Why I started building MeetingIQ.

I've been in meetings where someone makes a confident claim and the important question arrives too late. I wanted a way to catch those moments while the conversation was still happening.

The Windows desktop prototype works with user-selected meeting audio and surfaces possible questions. That's different from this website's scripted, fictional examples. The public demo does not listen, record, transcribe, or contact an AI service.

The tradeoff: a useful question can help a discussion, but a suggestion isn't evidence. It still needs a person to decide whether to ask it, and to verify the answer.

I'm still experimenting. The desktop package is not available for download here while distribution and data-handling questions are being reviewed.

01 / THE MOMENT

“We assume it's ready.”

A confident statement can hide an unanswered question.

02 / THE PROMPT

What needs checking?

Surface the missing evidence, owner, or condition before the discussion moves on.

03 / YOUR CALL

You decide what to ask.

Review the suggestion, use your judgment, and keep the conversation yours.

MEETINGIQ / INTERACTIVE EXPERIENCE

Spot the moment that matters.

SYNTHETIC MEETING · NO AUDIO CAPTURE

Play through a fictional project meeting. See how a question-first companion could surface assumptions, risks, and decisions while the conversation unfolds.

FICTIONAL TRANSCRIPT

Project launch review

QUESTION-FIRST ANALYSIS

Waiting for a moment…

Select “Next moment” to start the sample meeting.

This is a scripted demonstration, not the live MeetingIQ engine.

All people, statements, and decisions in this demonstration are fictional. It does not listen, record, transcribe or make API calls. The real desktop app has separate setup, provider, and data-handling requirements.

THE DESKTOP APP

Beyond the browser sample.

01

Question Radar

Follow the discussion and surface a timely question. Keep earlier suggestions in history, then copy, save, or export the ones you want.

02

Answer Assist

Paste reference notes before listening. Get draft answers with source excerpts and missing details to check. Matching an excerpt does not verify the facts or the answer's interpretation. Review every suggestion before using it.

03

Context that fits

Choose a saved professional profile, starting with General participant if you prefer. Your role helps shape the questions.

04

Control the session

Choose the audio source, pause either analysis panel, and export useful answers. Transcription continues while panels are paused. Stop listening ends audio capture.

FROM IDEA TO DESKTOP PROTOTYPE

Built around the conversation.

MeetingIQ started with a practical question: can AI help someone ask a better question during a meeting? Question Radar remains central. Answer Assist adds reference-backed draft replies when you have source material to work from.

The desktop prototype is a Windows portable app used through desktop Chrome. It needs Node.js 20+, npm, and provider setup. It is not an iPhone or Android app.

OpenAI mode is paid and sends captured audio, recent transcript, selected context, optional meeting context, and reference notes to OpenAI. Local processing needs separate Whisper, FFmpeg, and Ollama setup and does not automatically fall back to OpenAI. Running the interface on your computer does not automatically make the processing local.

Pause view holds the display while processing continues. Pause radar and Pause answers stop new analysis for that panel, but audio transcription continues and may still incur charges. Use Stop listening to end capture. Use meeting audio only with the required permissions.

Meeting history and pasted references stay in the browser session unless exported; refreshing clears them. Saved profiles and API settings are stored under your Windows account. Review exports before sharing, because they can contain transcript excerpts.

KEEP EXPLORING

One question. A whole project.

There are a few more rabbit holes in the workshop.

Explore other projects ↗