Skip to content

FROM CONVERSATION TO ACTION

Spoken plans.
Saved tasks.

Review the slides today. Reply tomorrow.
Talk to an AI character to add, complete, or defer tasks.

Check the changes, save them,
and pick up where you left off.

Try tasks — no account ↗Watch the 45-second demo ↓

Text tasks are free. Voice: 3 minutes after verified sign-in.

THE LISTENING ROOM01 / 03

AI MEETING

The materials review is done. The email reply has been deferred to tomorrow.

VRoid B Available in the app

Open the VRM preview ↗

Hear where your tasks stand0:00 / 0:14

AFTER THE CONVERSATION

Review materials Done

Reply to email Deferred to tomorrow

Recorded in JapaneseAbout the recordings

01 — Clear your head

Turn “later” into
a clear next step.

Say what needs doing, then check the changes. Tasks stay in this browser, ready when you return.

IN THIS SAMPLE

YOURead the current record and briefly tell me the status of both tasks.

AIThe materials review is done. The email reply has been deferred to tomorrow.

These two demonstration tasks were updated in the real session. They are not connected to an external task service.

SEE IT IN USE

Say it. Check it.
Keep the result.

From adding tasks to interrupting and confirming changes.
See the complete flow in a 45-second demo.

Recorded demonstration · 45 seconds · captionsWatch on YouTube ↗8-second TikTok preview ↗

Try it with your own tasks.
The voice, task validation, and persistence code is open source.

Try tasks ↗Explore on GitHub / Star ↗

Recording conditions and what has been verified

Samples pair synthesized user speech with real Gemini Live responses. They are not recordings of a person speaking into a physical microphone. Interview and English samples use the repository personas; the wait before each response is shortened to 0.6 seconds.

The 14-second task sample and 45-second film come from a real session in a dedicated recording layout. The script confirms only proposals matching the authored input. Task results and AI responses are not invented.

These recordings do not establish human-perceived naturalness, long physical-microphone sessions, or quality received by other Meet / Zoom participants.

Read the validation notes

YOUR TURN

Now, make it yours.

Start with a task, or try talking it through.
Choose the way that works for you.

01

Write down one next step.

Add, complete or defer tasks. Save them in this browser and export a backup.

Free · no registration · no automatic cross-device sync

Open your tasks ↗

02

Have a conversation.

Verify your email to try the voice experience. The app asks for microphone access when you start a conversation.

3 minutes · once per UTC day · 30 shared sessions/month · no automatic billing

Try a voice session ↗

YOURS TO BUILD ON

Your environment.
Your kind of conversation.

Adapt practice scenarios, characters and voice providers. Explore the source and set it up in your own environment.

Explore the code on GitHub ↗Download the OSS preview ↗

Extract the preview archive and run it with Python; no Node.js setup or API key is needed.

If it looks useful, star the project to follow its progress.

Star on GitHub ↗

Quick start Validation & limits Releases Apache-2.0

git clone https://github.com/FORIFOR/AI-meeting.git
cd AI-meeting
corepack enable
corepack pnpm install --frozen-lockfile
corepack pnpm dev:oss

Node.js 22 recommended. These commands start the preview. AI conversation requires provider configuration. Model assets have separate terms.

BUILT, TESTED, MEASURED

Your team's next step.

Turn a conversation into shared tasks. Start with a small team of up to five people.

PRODUCT INTRODUCTION

See the workflow in action.

A 45-second demo with real Gemini responses. The task interface passed ten repetitions, including save, restore, and failure handling.

Demo input is synthetic. It does not represent human adoption or satisfaction.

LIMITED PILOT

Technical preparation is complete.

Verified-email invitations, shared tasks, administrator controls, deletion and restoration were checked through an automated workflow on the real cloud service.

Scope: five members, 30 days, three-minute voice sessions, non-confidential tasks. Paid pilot terms and data requirements are agreed separately.

OPERATIONS

Prepare for failure, too.

Usage limits, datastore monitoring and seven-day recovery retention are configured. Failed checks block promotion, and failed post-release checks trigger a rollback.

This limited configuration is live. This is not a claim of enterprise SSO, meeting integrations, an SLA, or unrestricted production readiness.

1,100 automated tests passed
September 14, 2026 · 124 files. The limited team workflow separately passed 14 cloud checks; the 100-human-conversation milestone remains unmet.

Open a team workspace ↗ Evidence & operating scope ↗ Discuss paid setup ↗

A CONVERSATION FOR YOUR TEAM

Voice AI for
your workflow.

Turn conversations into tasks or rehearse an interview or sales call. Discuss workflow design, voice AI implementation and deployment support.

Task planning and scenarios for your roles
Character and voice configuration and terms
Custom implementation, deployment and operational support

Availability, scope, timing and fees are agreed individually. Submitting this form creates no contract or payment.

Confirm the use case, audience, and constraints
Propose scope, validation, timing, and fees
Start implementation or customization after agreement

GOOD TO KNOW

Before you begin.

Where do tasks and voice data go?

Typed tasks stay in this browser. A voice session sends audio, speech and the tasks it uses to Google’s voice AI. Sign-in uses your email; the server retains usage-limit records.

What does this website measure?

This site counts sample starts and completions, scenario switches, GitHub and quick-start clicks, and accepted inquiries by day. No cookies or visitor identifiers are used. Conversations, recordings, tasks, form messages and emails are excluded from analytics. GitHub clicks are not stars. We respect Do Not Track and Global Privacy Control.

Are business inquiries public?

No. Only the operator can access the reply address and inquiry for follow-up. It is not sent to an AI model or advertising analytics. A deletion expiry of 90 days is set when stored. No automatic email is sent.

Is Meet or Zoom ready to use?

Meeting integrations are experimental and need provider-specific setup and call validation. Their scope differs from the public voice experience.

Is there a subscription plan?

This beta does not offer subscriptions or cross-device task sync. Task management and voice sessions within the shared allowance are free. Implementation services are scoped individually; self-hosted cloud providers charge under their own terms.