Private Beta · Mac

One chat box. All your subscriptions.

Claude Code and Codex under one roof. Maestro routes each task to the right engine and writes straight to your real files. Your machine. Your code never leaves it.

Runs 100% on your computer. No API keys, no cloud.
MMaestro
Maestro
ChatReview

Start a local run

Select a working directory, choose Auto or a specific engine, then send a prompt. Runs stay tied to this Mac and the selected folder.

codexo-next
AutoClaudeCodex
Options
Send a prompt... (Enter to send, Shift+Enter for newline)Send
Works with
Claude CodeCodexElevenLabsGemini (soon)Grok (soon)
$0 per token

Backed by

ElevenLabs Grants

Argus, Maestro's voice layer, is built on an ElevenLabs voice agent under the ElevenLabs Grants program.

The gap

Two terminals. Zero shared memory.

You pay for Claude and you pay for ChatGPT. Then you spend your day alt-tabbing between two terminals that have never heard of each other. Same context pasted twice. Same repo explained twice. Two engines, each doing half a job, with you as the glue.

Without Maestro

Two terminals, two histories, no shared memory

You copy context between engines by hand

Metered API bills if you want automation

Agents are a demo you watched, not a team you run

With Maestro

One cockpit, every engine, your real files

Each task routed to the engine that does it best

Loops run on the flat subscriptions you already pay

Talk to Argus and the work starts

Meet Argus

Talk to your computer. It gets to work.

Argus is the voice inside Maestro. Say what you need and Argus does it: routes the task to the right engine, opens the app, clicks the button, types the reply, reports back when it's done. Built on an ElevenLabs voice agent, wired into everything Maestro can do: your engines, your files, your screen.

This is not a smart speaker that sets timers. It's a hands-on operator for your Mac. You talk, it acts, you watch it happen.

Say the task once. Argus picks the engine and runs it.

It can see your screen and drive your Mac. Click, type, navigate.

Ask what happened while you were out. It tells you, out loud.

See the roster it commands

Argus voice runs on ElevenLabs. Your code and files stay on your machine. See Privacy.

Computer use

It doesn't just answer. It operates.

Maestro can use your Mac the way you do. It sees the screen, moves the cursor, clicks, and types: native control, not a browser trick. And when the job lives on the web, Maestro drives its own bundled browser, so agents can fill forms, pull data, and finish web tasks end to end. The browser side works from anywhere, including your phone.

Native Mac control

Screen awareness plus real input. Maestro handles the apps you actually work in.

Bundled browser

Its own browser for web work. Kick off a task from your phone, watch it finish from anywhere.

Real-world loops

Your operations, on a schedule. On subscriptions you already pay.

Loops are agents that run your actual day on a cadence: a morning brief before you're up, inbox triage three times a day, pipeline review every morning, an end-of-day wrap. They run on your flat Claude and Codex subscriptions. No metered API bills, no per-token anxiety. Set the cadence once and the work happens whether you're at the desk or not.

Morning brief

07:00 daily

What matters today, ready before you are.

Inbox triage

3 passes a day

Sorted, drafted, flagged, before it piles up.

Pipeline

09:00 daily

Every open deal reviewed, next actions surfaced.

End of day

17:30 daily

What shipped, what stalled, what's queued for tomorrow.

Oversight

Watch the work. Approve the output.

Autonomy without visibility is a liability. Maestro gives you two views: the Work dashboard, where you watch agents run in real time, and the Review dashboard, where everything they produce waits for your sign-off before it ships. Nothing goes out the door because a model felt confident. You stay the last gate.

Work dashboard

Live view of every running agent: what it's doing, on which engine, right now.

Review dashboard

Every deliverable queued for your review. Approve, edit, or send it back.

The roster

The agents we run our own business on. Included.

Maestro ships with the same agents Codexo uses to operate daily, not demo agents built for a screenshot. A web development agent that plans and ships. A VP of Sales agent and a marketing agent built on the playbooks popularized by Alex Hormozi and Russell Brunson: offers, follow-up, pipeline discipline. You get the roster on day one, and you can rewrite any of them in plain markdown.

Web Developer

Scopes, builds, and ships against your real repos.

VP of Sales

Pipeline reviews, follow-up sequences, deal hygiene. Trained on the playbooks popularized by Alex Hormozi.

Marketing

Offers, copy, and funnels built on the frameworks popularized by Russell Brunson.

Agent playbooks are inspired by publicly available frameworks. No affiliation with or endorsement by Alex Hormozi or Russell Brunson.

How it works

Three steps to a working orchestra.

01 / Install

One Mac app. No API keys, no cloud accounts to create.

02 / Connect your CLIs

Maestro finds the Claude Code and Codex you already have and plugs them in.

03 / Say the word

Type it or tell Argus. Maestro routes it, runs it, and shows you the result to approve.

The film

Stop paying per token.

Seventy-two seconds on the whole idea: agents on schedule loops, running on flat subscriptions instead of metered APIs, on a machine you own. Watch it before you scroll further.

Agents on schedule loops, running on your subscriptions, on a machine you own.

Privacy

Your keys. Your machine. Your code.

Maestro is not a cloud service wearing a desktop icon. Your engines run locally through the CLIs you already installed, under the accounts you already pay for. Your repos, files, and prompts to the engines stay on your Mac. We don't proxy your traffic, we don't store your code, and we can't read your work.

One honest exception: Argus speaks and listens through ElevenLabs's voice service, so voice audio travels to ElevenLabs to be understood and spoken back. Your code and files are not part of that stream. Voice is optional. Everything in Maestro works by keyboard.

Data flow

Your Mac

Your prompt

Maestro (local)

Your CLIs (local)

Your files (local)

Everything above stays on this machine

Argus voice (optional)

Audio only, to and from ElevenLabs. Never your code or files.

Codexo servers
(zero access to your code and prompts)

Engines

Built for the engines you already pay for.

Maestro routes each task to whichever engine does that job best. When new engines land, your setup gets stronger without changing anything.

Live

Claude Code

Anthropic

Deep reasoning and long-horizon builds.

Live

Codex

OpenAI

Fast, capable, and relentless on code.

Soon

Gemini

Google

On the roadmap.

Soon

Grok

xAI

On the roadmap.

Roadmap

What's live, what's next.

Now

Live in private beta

The cockpit: one chat box, multi-engine routing, real file writes

Argus voice assistant (ElevenLabs voice agent)

Native Mac computer use and bundled browser use

Scheduled real-world loops on flat subscriptions

Work and Review dashboards

The agent roster: web dev, VP of Sales, marketing. All markdown-defined and editable

Next

In build

Public Mac build (funded by founder deposits)

Gemini and Grok engines

Later

On the horizon

Shared memory across every engine

Beyond Mac

Founder access

Get in before launch. Lock your price for good.

Maestro is in private build for Mac. Join the waitlist free and you're first in line when the Mac app ships. Or put down a deposit: it funds the public build and locks a discounted lifetime monthly rate when your invite opens. Change your mind before launch and we refund it, no questions.

Waitlist

Free · Private beta for Mac

Be first to get the Mac app

Founder Discord access

By joining, you agree to receive automated and AI-generated updates from Maestro and accept our Terms and Privacy Policy. Consent is not a condition of purchase.

Founder deposit

Refundable before launch · funds the public build

Selected deposit: $50. Lifetime access at $15/month when your invite opens.

You already pay your AI providers directly. Maestro is the tool that makes them work as one. We never charge for tokens or take a cut of your usage. At launch: Free tier and Pro at $30/mo.

Community

Build alongside the people building it.

The Discord is where the beta actually happens: build logs, loop configs, agent markdown swaps, and direct access to the founder shipping it. Waitlist and deposit members are both in.

Join the Discord

FAQ

Questions, answered.