The desktop app runs it. Your phone can too.
Buttonizer is a button panel for your computer. Install the desktop app, and your AI agent writes the commands — or you make buttons yourself. Use them right on your computer. Then add your phone as a remote and tap the same buttons from anywhere.
macOS today · iPhone beta on TestFlight · Windows & Linux coming soon
The guide is raw Markdown written for your AI agent — you hand it the URL, you don't read it like a manual.
A script to rename files. A tiny site to check a status. A tool to restart a service. You don’t need another app — you need a button. If there is something you do over and over, turn it into a button once and stop repeating yourself.
Tell your coding agent what you want. It writes the command.
The agent adds it as a button on your computer. No code to review.
Press the button on your phone. The job runs on your computer.
The desktop app is where buttons live and run. Your phone comes later, as a remote.
macOS today — Windows and Linux are coming soon. It watches your button folder and runs jobs on your machine — nothing runs in the cloud.
Ask your agent to make them (point it at buttonizer.app/guide), or add them in the app.
Click a button in the app or from the tray. Watch live output and keep per-button history.
“Oh — it works on my phone too.” Pair once and the same buttons show up on your phone, so you can run them from anywhere.
Three steps, no app development.
“Read buttonizer.app/guide and make me a button that resizes every screenshot in my folder.”
“Combine your agent skills into one button.”
Include buttonizer.app/guide so it follows the exact button format.
“Resize every screenshot in my folder.” The agent writes the command and creates the button.
Press the button on your phone. Get a notification when the job finishes.
If you use a skill again and again, make it a button — anything you can run in a terminal becomes a one-tap action.
Shrink a folder of screenshots to a fixed width.
“Resize every screenshot in my folder.”
Ship the current branch to staging and get pinged when it’s live.
“Deploy my app to staging.”
Copy a project or photo library to an external drive.
“Back up ~/Projects to my NAS.”
Ping services and report disk, CPU, and uptime.
“Check my server and tell me what’s wrong.”
Move old files, empty the trash, keep the folder tidy.
“Clean up my Downloads folder.”
Bounce a dev server or Docker container from your phone.
“Restart my dev server.”
Buttonizer is a remote control, not a cloud runner.
Everything you need to run jobs you trust.
Shell scripts, CLIs, builds, deploys — anything your computer can run.
Get a push on your phone the moment a job completes.
Watch stdout/stderr while it runs, and keep history per button.
Trigger from your phone, the desktop app, or your own HTTP call.
Your agent turns its own skills into a single button — one tap runs the whole flow.
Organize buttons as your collection grows.
Run in the background, stream logs, or keep a service alive.
One small button.yml covers all of this.
short · long (background) · service (keep alive){{placeholders}} filled from parameters at run timeThe desktop app is free — install it on as many computers as you like. You only pay when your phone connects to more than one of them at the same time.
In plain words: the desktop app never costs anything. The subscription is only for using your phone as a remote for more than one computer. Billed monthly through the App Store — cancel anytime.
Nothing to run buttons on your computer — the desktop app is free on every machine. You only subscribe when you want your phone to control more than one computer.
Yes — buttons run on your computer, so the desktop app is required. The phone app is optional: install it later to run the same buttons from anywhere.
A capability your AI agent (or harness) has — reading files, running a build, searching the web, and so on. Point the agent at the guide and it can turn a skill, or several skills combined, into a button.
An AI coding agent or harness that runs on your computer (for example, a coding assistant you already use). You describe the task; it writes the command and creates the button.
No. If you can describe the task, the agent can make the button. Power users can edit the command directly.
Your commands, their output, and your files stay on your computer. We only keep pairing identifiers, push tokens, and subscription records. See the Privacy Policy.
The desktop app runs on macOS today (Windows and Linux are coming soon). The iPhone app is coming soon to the App Store.
Point your agent at buttonizer.app/guide and describe the task. The agent writes a button.yml and it appears in the app.