Developer-first · Voice-controlled
Run your Mac and AI coding agents by voice.
Conduct AI coding agents, folder-backed workspaces, adaptive layouts, common Mac apps, public-browser workflows, files, and media by voice. Sensitive actions ask first, and Vochestra reports an observed result when a tool can verify it.
or request a complimentary 30-day trial·Watch the launch film
Actively developed — supported workflows, execution precision, and reliability improve release by release, with updates typically shipping weekly during early access.
Voice command to a coordinated Codex session on the Vochestra canvas.
78 seconds. See agents, folder-backed workspaces, and adaptive layouts move as one system.
One conductor for every project stage. Keep your hands on the work.
Global hotkey (⌥Space) anywhere in macOS. Gemini Live voice with one consistent persona. On supported audio routes, when on-device Speech and verified output isolation are available and permitted, you can interrupt it mid-sentence by speaking.
Claude Code, Codex, Gemini CLI, OpenCode, Cursor and Grok Build run in glass terminals on an infinite canvas. Agents get codenames — “Nova is on it.”
On supported public web pages, follow-ups like “open the first link” use the page’s exact link target and continue in the same canvas browser.
Ask “what’s this error?” and Vochestra looks at your screen silently. Or capture a screenshot straight to the canvas.
Rectangles, arrows, freehand, text — annotate screenshots by voice or hand, then attach them to a supported terminal agent.
For example, “paste that into Slack” can type into an accessible text field in the frontmost app, with a visible lightning flash over the window.
When on-device speech recognition is available, captions stay local. Otherwise the app uses the Gemini conversation transcript as a network fallback.
Ask for a track and Vochestra can use its supported Spotify or public-web search tools — then report whether playback was verified. Apple Music controls cover current playback.
Hard questions can use a configured Claude Code or Codex CLI in a restricted, non-interactive mode under the subscription you already manage.
Each workspace is bound to a folder and keeps its own terminals, browsers, captures, and saved canvas. Switch, rename, create, or remove the Vochestra workspace by voice without deleting the project folder.
Ask for an equal grid, two columns, a vertical stack, or Focus + Stack with one named agent made primary. Vochestra resolves the target and preserves the rest of the working stage.
Buying the license is step one. The app unlocks after activation and a Gemini key. Full setup guide.
Install Vochestra on an Apple Silicon Mac running macOS 15 or later. Enter the email and license key from your purchase or trial email.
Voice and reasoning use your Google Gemini API key. Google bills that usage. Then grant Microphone access so you can talk.
“Open a Claude terminal and fix the failing test.” Vochestra routes the request to bounded tools, then reports an observed result or says when verification is uncertain.
Your Gemini key is stored in macOS Keychain.
On-device speech captions when supported.
Local private-data files use owner-only permissions.
Ambiguous or sensitive actions ask first.
A perpetual individual license — not a subscription. Personal and Pro include a limited update window; covered versions keep working after that window ends.
Early-access pricing for the current paid plans.
Use
Personal · Personal projects only — not paid work
Pro · Work and paid-project use
Founder · Work and paid-project use
Macs
Personal · 1
Pro · Up to 3, used only by you
Founder · Unlimited, used only by you
Live terminals / Mac
Personal · 1 at a time
Pro · Up to 5
Founder · Uncapped by Vochestra
Workspaces
Personal · 1
Pro · Up to 3
Founder · Unlimited
Updates
Personal · 1 year
Pro · 2 years
Founder · Lifetime desktop
Hobby use · One Mac
Hobby only — not for client work, employment, or paid projects.
Workspace capacity
1
One complete project stage
Work and paid projects · Three Macs
Workspace capacity
3
Move between active projects
Maximum scale · Lifetime updates
Workspace capacity
Unlimited
A project stage for everything
You bring your own Gemini key and coding-agent subscriptions. Personal is hobby-only. Refunds · FAQ · 30-day trial
Pay once and keep the versions covered by your purchase. Personal is for personal projects on 1 Mac with 1 agent or shell terminal running at a time, with 1 workspace. Pro adds work and paid-project use, up to 5 terminals at once per Mac, 3 workspaces, 3 Macs, and 2 years of updates. Founder lets you run as many terminals as your Mac and providers allow, supports unlimited workspaces and Macs used only by you, and includes lifetime desktop updates. Refunds and statutory rights are described in our refund policy.
Every live Vochestra agent or shell terminal counts toward the per-Mac parallel limit. Every plan includes every supported agent provider. Every plan also includes Auto Grid, Focus + Stack, columns, stacks, and manual layout, all controllable by voice. Actual capacity and availability depend on your Mac, installed agents, provider plans, and provider limits.
You bring your own Google Gemini API key and coding-agent subscriptions. Their providers control availability, limits, and billing; Vochestra never adds an AI-usage charge.
Vochestra controls common Mac apps, public-browser workflows, files, media, and supported AI coding agents by voice through 60+ bounded tools. Sensitive or ambiguous actions ask first. When a tool can observe the outcome, Vochestra reports the observed result; otherwise it says when verification is pending or uncertain.
Yes. Every paid plan supports Claude Code, Codex, Gemini CLI, OpenCode, Cursor Agent, and Grok Build in glass terminals on an infinite canvas. Each agent gets a codename, and you delegate work by voice. Personal runs 1 agent or shell terminal at a time, Pro runs up to 5 in parallel per Mac, and Founder lets you run as many as your Mac and providers allow. Those agents must already be installed on your Mac; Vochestra does not include their subscriptions.
For agent workflows, yes. Vochestra launches and coordinates the supported tools you already have. You can still use voice for supported Mac, browser, file, and media workflows without those agents.
No. Vochestra is a one-time license — Personal $49, Pro $99, or Founder $149. You bring your own Google Gemini API key, and Google bills any usage under your account and terms. Token counts shown in the app come from the API; cost uses an exact published model rate when available and labels its closest-rate fallback as an estimate. There is no monthly fee from us.
Create a key in Google AI Studio using Google's Gemini API key guide. Google — not Vochestra — sets availability, rate limits, and billing for that key. The app will not finish onboarding without a working key.
Approved applicants receive a complimentary 30-day Founder trial with commercial-use rights, unlimited workspaces, unlimited Macs used only by the trial holder, and as many terminals as their Mac and providers allow. The 30 days start on first activation. No payment method is required, there is no automatic charge, and the trial expires automatically. Purchase Personal, Pro, or Founder to continue after it ends.
Personal is for one person's personal projects on 1 Mac, runs 1 agent or shell terminal at a time, includes 1 folder-bound workspace, and includes updates released during the first year. Pro adds work and paid-project use, runs up to 5 terminals in parallel per Mac, includes up to 3 workspaces, activates up to 3 Macs used only by that person, and includes 2 years of updates. Founder provides unlimited workspaces, lets that person run as many terminals as their Mac and providers allow, activates unlimited Macs used only by them, permits work and paid projects, and includes lifetime desktop updates. Covered Personal and Pro versions keep working after their update window ends.
No. Personal is an individual non-commercial license. If you use Vochestra at work, for freelance, or on paid projects, buy Pro or Founder.
Current Vochestra builds require an Apple Silicon Mac running macOS 15 (Sequoia) or later. Intel Macs and older macOS versions are not supported. Update access is included for the period listed with your plan; we do not promise unattended installation.
Microphone is required for voice. Screen Recording, Accessibility, Speech Recognition, and Music app control unlock the matching workflows. The setup guide walks through each prompt.
After Creem confirms payment, we email the license key to the checkout address. A payment receipt is not the key. Then download the app and activate with that same email. If nothing arrives in about ten minutes, check spam and contact support.
Email us from the purchase address and we will help you move the license. You may not share the key. Each activated Mac must be used only by the license holder.
The voice conversation sends audio directly from your Mac to Google Gemini under your API key. If you ask Vochestra to inspect or act on content, the needed screen, file, terminal, browser, clipboard, contact, message, note, or context data may also be sent to Gemini. On-device captions are used only when your Mac supports them; otherwise Gemini's network transcript is the fallback. The Vochestra operator does not receive those raw contents unless you explicitly send a report or message.
Vochestra is built around multi-step developer workflows: it runs supported coding agents, works with public web pages in its canvas browser, types through macOS Accessibility where supported, sees and annotates your screen when asked, learns bounded recent habits, and can route hard questions to a configured Claude Code or Codex CLI.
Optional · reviewed in batches
Licenses are on sale now. If you would rather try first, request an invite. Approved users receive a complimentary 30-day Founder trial with work and paid-project use on unlimited Macs used only by them. It has no Vochestra terminal cap, starts on first activation, expires automatically, creates no automatic charge, and requires a paid plan to continue.