Meow is in public beta

Decision Cue

Strategy, made tangible.

An independent product lab turning real work problems into software, AI systems and experiments.

Explore the work
STRATEGY · AI · PRODUCT DESIGNFOCUS · DECIDE · SIMULATE · ACT

ABOUT DECISION CUE

Thinking is only the start.

The interesting part is what happens when an idea has to become a system, an interface and something another person can actually use.

Decision Cue explores that transition through independent product experiments across focused software, multi-agent AI, workflow simulation and voice automation.

Some become public products. Some stay in testing. Others remain experiments. Each one is built far enough to reveal what works and what should come next.

FOCUS → DECIDE → SIMULATE → ACT

Four directions. One principle: make the thinking real.

SELECTED BUILDS

Products & experiments

One product shipped.
More in active development.

FOCUS · DESKTOP SOFTWARE

Meow

Public beta
Explore Meow

Your own space to browse.

Windows 11 · 64-bit · Public beta

A Windows browser designed around focused knowledge work, with system audio input, supported screen-sharing privacy and a floating webpage view. Built and released as a downloadable Windows 11 public beta.

Desktop App · Browser UX · Privacy · Localization

What it does

Bring system audio into compatible web apps, keep the Meow workspace out of supported screen-sharing tools, and float the part of a webpage you need above your work.

When Meow is closed normally, website sign-ins and temporary site data are cleared while settings, bookmarks, downloads and imported extension data remain available.

See the Meow product page for downloads, technical details and terms.

Illustration of Meow sending audio to a webpage and floating a panel above a laptop.
Product illustration · Public beta

DECIDE · MULTI-AGENT AI

Stage One

Under testing
Discuss Stage One

Turn a question into evidence, options and challenge.

Agents: testing · Rehearsal Universe: R&D · Not public

A multi-agent consulting environment designed to frame complex questions, research evidence, test assumptions and challenge recommendations. Rehearsal Universe extends the concept into scenario-based decision rehearsal.

Multi-agent AI · Research Orchestration · Decision Support · Scenario Reasoning

Inside Stage One

The first layer is designed to use specialized AI agents to structure a problem, gather evidence, develop alternatives and challenge emerging conclusions.

Rehearsal Universe explores a second layer: representing stakeholders, scenarios and consequences so that decisions can be examined before they are made. This layer is in R&D.

Currently under testing. Not publicly available.

Preview of Stage One’s AI consulting agent workflow and conversation.
Testing preview · Under testing

SIMULATE · WORKFLOW MODELING

SimUI

Under testing
Discuss SimUI

Describe the workflow. Test the change.

Under testing · Not publicly available

A workflow simulation platform designed to turn visual or natural-language process descriptions into models, then explore capacity, bottlenecks and alternative operating designs before changes are made in the real world.

Natural-language Modeling · Simulation · Operations · Scenario Testing

Inside SimUI

Build a workflow visually or describe the process through reviewed natural-language proposals. SimUI turns that logic into a model that can be tested under different demand, capacity and operating assumptions.

The platform explores how to compare scenarios, identify bottlenecks and examine changes before touching the real operation.

Currently under testing. Not publicly available.

SimUI interface preview showing a bank staffing workflow model.
Testing preview · Under testing

ACT · VOICE AI

CallRunner

Planned
Discuss CallRunner

Delegate a phone task without giving up control.

Concept development · Not available yet

A planned voice-agent workflow designed to clarify the user’s goal, confirm permissions, make the call, manage retries and exceptions, allow human intervention and record the outcome.

Voice AI · Agent Workflows · Tool Use · Human-in-the-loop

The plan

CallRunner is designed around controlled delegation rather than unrestricted autonomy. The planned agent would first clarify the task and authorization, then act within the agreed scope.

When information is missing or an exception occurs, the planned workflow would return to the user for clarification or intervention before continuing.

Planned capabilities. Not yet publicly available. The concept expects users to bring their own AI and telephony accounts.