Safeguard on Mac, iPhone and iPad: one setup, step by step

Safeguard runs across Apple devices as one platform. On a Mac you get the desktop app and the Safeguard CLI in Terminal. On iPhone and iPad you get the App Store app, which includes a Safari extension. Claude and ChatGPT connect through the Safeguard MCP server. Sign in once on each device and the same findings follow you.

You also keep what you already pay for. Bring the Claude, ChatGPT or Codex plan you already pay for and use it inside the Safeguard CLI. Safeguard adds no extra charge for that usage; your plan's own usage limits and terms still apply. And if you prefer to stay in the chat you use every day, connect Claude or ChatGPT to the Safeguard MCP server and ask about your vulnerabilities from there.

This guide walks through the whole Apple setup in order: Mac first, because that is where the work happens, then iPhone and iPad, then the chat assistants. Every device fact here comes from the Safeguard download page as checked on 21 October 2026, and every CLI fact comes from the CLI docs.

What runs where across your Apple devices

Each surface reaches the same account and the same data. You are not choosing one; you are choosing which door to use at a given moment.

Three things in that table are easy to miss.

First, there are two Mac desktop builds. The download page tells you to check your chip under the Apple menu, About This Mac. If it says "Apple M" followed by a number, you want the Apple Silicon build. If it says Intel, you want the x64 build linked just below it.

Second, the Safari extension on Mac is still in Apple review. On iPhone and iPad it ships inside the app today. If you browse on a Mac and want the extension now, the Chrome Web Store version covers Chrome, Brave, Arc, Vivaldi, Opera and Edge, and there is a Firefox version too.

Third, nothing on iPhone or iPad needs a separate account. The app signs in to the same Safeguard workspace you use on the Mac.

Step by step: set up the Mac

The Mac is where you write and fix code, so set it up first: the desktop app for the full platform in a native window, and the Safeguard CLI for Terminal.

1. Check your chip

Open the Apple menu and choose About This Mac. Note whether it lists an Apple M chip or an Intel processor. This decides which desktop build you download. The CLI installer works this out for you, so this step only matters for the desktop app.

2. Install the desktop app

Go to safeguard.sh/download and pick the Mac build. The Apple Silicon file is a DMG for macOS 12 or later. Intel Macs use the x64 build linked from the same section.

The download page says the Mac build is signed with Safeguard.sh Inc's Developer ID and notarized by Apple, so there are no warnings to click through. Open the DMG, drag the app to Applications, and launch it.

3. Sign in and learn the two halves of the window

The desktop docs describe the window in two parts. On the left is the web app, and you sign in to it exactly as you would in a browser. On the right is a local panel with tabs for Scan, Terminal, MCP, Tasks, Files and Settings. Press Cmd+Shift+S to show or hide that panel.

The panel is deliberately fenced. Its Terminal and Files tabs only work with allowlisted programs and folders you have granted. When a task in the panel is about to take a destructive step, it raises an approval request and waits for you.

The left half is where you "ask your supply chain anything" in plain English, which is how the download page describes the desktop app. The right half is where local work happens.

4. Install the Safeguard CLI

Open Terminal and run the installer:

curl -fsSL https://cli.safeguard.sh/install | bash

The download page says the installer detects your operating system and verifies the checksum. You get a single static binary, version 1.4.0 on the download page, with nothing else to install. There is no Node, Python or JVM to set up first.

The desktop docs note that the app bundles the CLI with its scanner and its Terminal tab allowlists safeguard, so you can use the CLI from macOS Terminal or from the desktop panel.

Do the work properly with the Safeguard CLI

The desktop app is a good place to look across everything. The CLI is where you get things done on your own machine. It scans code, container images and SBOMs locally, sends only the results to the platform, and includes an assistant that can fix what it finds. The assistant writes code, not only scans it.

Install and first session

You already ran the installer above. Now move into a project and start the CLI with no arguments:

cd ~/code/payments-api
safeguard

That opens the interactive terminal. Type /login and your browser opens so you can approve the session. Then scan the folder you are in:

/scan .

Once you are signed in, the local runner also picks up scan jobs scheduled from the platform. When it runs those jobs, your source code stays on your Mac. Only SBOMs and findings are uploaded.

Ask in plain language

Anything that does not start with a slash goes to the assistant. The docs give three examples, and they show the range well:

what's the latest file in ~/Downloads
how many critical vulns do we have
fix the lodash CVE in this repo

The first is an ordinary question about your Mac. The second reads your Safeguard data. The third changes code. The assistant has tools for your local files and tools for your Safeguard data, and it picks whichever the question needs. Type @ to attach a file from the folder you started in, and use /resume tomorrow to pick up the last session in that same folder.

Writing and fixing code

Because the same assistant can see your findings and your files, you can go from "what is critical here" to "fix it" without switching tools. Writing a file, editing code and running a command all ask before they act. A typical loop is: ask for the fix, read the proposed edit, approve it, ask it to run your tests, then /scan . again to confirm the finding is gone.

This walkthrough shows the same loop driven from Search: attach a computer running the Safeguard CLI and a folder, ask for a fix, review the diff and approve the command before it runs. The demo itself notes that computer work from Search is available and expanding.

Click through it at your own pace, or open it full screen.

Use your Claude, ChatGPT or Codex plan

This is the point that makes switching easy. Bring the Claude, ChatGPT or Codex plan you already pay for and use it inside the Safeguard CLI. Safeguard adds no extra charge for that usage; your plan's own usage limits and terms still apply.

/model shows or switches the model you are using. For connecting a model, see the CLI docs and the /model picker. You keep the subscription you chose and gain an assistant that also knows your vulnerabilities.

Approvals

/permissions sets one of three modes:

• Ask prompts before every action.
• Auto approves known actions and asks about anything new. It is the default.
• Yolo approves everything.

Use Ask in a repository you do not know. Use Auto for daily work in code you own. Keep Yolo for a scratch folder where a wrong command costs nothing.

Scripts and CI with safeguard exec

The same binary works without the interactive terminal. These run from any shell on your Mac:

safeguard scan --dir .
safeguard scan --image nginx:latest
safeguard scan --sbom sbom.json
safeguard validate sbom.json
safeguard exec "summarise the critical findings in this repository"

safeguard scan builds an SBOM and matches it against vulnerability databases, writing JSON files such as vuln.json and license.json under ./output. safeguard validate checks CycloneDX, SPDX and VEX documents. safeguard exec runs one non-interactive assistant turn, which is what you put in CI. Policies and gates are evaluated in the platform, not by the CLI, so the CLI produces findings and the platform decides whether a build passes.

If another coding agent wrote the change, the CLI is still useful: scan what changed before you merge. We cover that flow in checking coding agent changes with the Safeguard CLI, and the wider assistant in using the Safeguard CLI as an AI assistant on your machine.

Step by step: iPhone, iPad and the Safari extension

With the Mac working, add your phone and tablet. The download page describes the mobile app plainly: the same answers on a phone. It is one app for both iPhone and iPad, from the App Store, and the store handles signing and updates.

1. Install the app

On iPhone or iPad, open the App Store and install the Safeguard.sh app. The download page links to it directly, so the simplest route is to open safeguard.sh/download in Safari on the device and tap the App Store link. That way you install the right app and not something with a similar name.

2. Sign in to the same workspace

Open the app and sign in with the same account you use on the Mac. The point of signing in to one workspace everywhere is that a finding you fixed on the Mac this morning shows as fixed on your phone this afternoon. There is nothing to sync by hand.

3. Turn on the Safari extension

The Safari extension for iPhone and iPad is included in the app. Installing the app does not switch it on. Do this once on each device:

1. Open the Settings app.
2. Tap Apps, then Safari, then Extensions.
3. Find Safeguard.sh and turn it on.

On iOS versions before 18, the download page gives the older path: Settings, then Safari, then Extensions.

Apple's Safari support page describes a "Share Across Devices" switch in the same Extensions screen. With it on, compatible extensions from your other devices show up so you can download them, and turning one on or off syncs across devices. Removing an extension still has to be done on each device.

4. Know what the Mac Safari extension status means

The download page lists Safari on Mac as in Apple review. Until that changes, do not expect to find it in the Mac App Store. On the Mac, use the desktop app, or the Chrome Web Store or Firefox version of the extension (1.0.4) if you browse in one of those.

5. Use the iPad as a second screen for review

On iPad, the app and Safari can sit next to each other. Apple's iPadOS 26 multitasking guide explains that you turn on Windowed Apps under Settings, Multitasking and Gestures, and then drag a second app to the left or right edge of a full screen app to place them side by side. A common layout is the Safeguard app on one side and the web interface in Safari on the other. Safeguard's web app docs say the interface is responsive for tablets and phones, and that mobile use focuses on review and alerts, such as reviewing and approving Griffin proposed remediations. Heavier work like editing policies or deep SBOM inspection is best done on the Mac.

A day across your Apple devices

In the morning, on the Mac, you run safeguard in the repo you are working on, /scan ., and ask the assistant to fix the critical dependency it found. You approve the edit, run the tests, and scan again.

At lunch, in the desktop app, you look across products to see whether the same package shows up elsewhere.

On the train home, on your iPhone, you open the app and check whether anything new came in. On the iPad that evening, you review a proposed remediation with the app on one side and the web interface on the other, then leave the code change for the Mac in the morning.

Start in the terminal, continue on the desktop, check findings on your phone.

Connect Claude or ChatGPT to Safeguard

The other way to keep what you already use is to bring Safeguard into Claude or ChatGPT. The Safeguard MCP server is hosted, so nothing needs to be installed on your Mac for this. You add a URL, sign in with OAuth, and approve access.

Claude on the web or Claude Desktop

Anthropic offers Claude as a desktop app for macOS and as an iOS app, according to its download page. Safeguard's setup guide covers Claude on the web and Claude Desktop:

1. Sign in to Safeguard at app.safeguard.sh, or create an account.
2. In Claude, open Customize, then Connectors, click Add, then Add custom connector.
3. Name it Safeguard Security and enter https://mcp.safeguard.sh/mcp/anthropic.
4. Under Authentication choose Sign in now or Sign in when needed, under OAuth client choose Register automatically, then click Add.
5. When Claude sends you to Safeguard, sign in, check the app name and the receiving host, and click Authorize access.

On Team or Enterprise plans, an Owner or Primary Owner adds the connector under Organization settings, Connectors, and members then click Connect. Anthropic's help article says custom connectors using remote MCP are available on Free, Pro, Max, Team and Enterprise plans, with Free limited to one custom connector. Safeguard's guide also says not to edit claude_desktop_config.json for Safeguard, because the server is hosted.

This walkthrough connects Claude to the Safeguard MCP server with OAuth and lets you choose which tools Claude can use.

Click through it at your own pace, or open it full screen.

ChatGPT

Safeguard's ChatGPT guide uses the general server address:

1. Sign in to Safeguard at app.safeguard.sh.
2. In ChatGPT, select the plus button, then Add custom MCP server.
3. Name it Safeguard Security and enter https://mcp.safeguard.sh/mcp under Connection.
4. Set authentication to OAuth. No client ID or key is needed.
5. Review the warning, continue, and create it.
6. Sign in to Safeguard when redirected and click Authorize access.

The guide notes that account and workspace policies apply, so if you do not see the option, ask your ChatGPT workspace administrator.

Which tools the assistant can see

Safeguard's MCP docs say a new workspace starts with six read-only tools for browsing products, projects, vulnerabilities and findings. An administrator turns on more tools one at a time. So a fresh connection can answer "what are our critical findings in the payments product" but cannot change anything until an admin decides it should.

Common mistakes

Downloading the wrong Mac build. Check About This Mac first. Apple Silicon and Intel use different desktop builds. The CLI installer picks the right binary on its own.

Looking for the Mac Safari extension. It is in Apple review. Use the desktop app on the Mac, or the Chrome or Firefox extension if you browse there.

Installing the iOS app and expecting the Safari extension to be on. It ships inside the app but stays off until you turn it on in Settings, Apps, Safari, Extensions.

Installing a different app with a similar name. Several unrelated apps use the word Safeguard. Follow the App Store link from safeguard.sh/download.

Asking about vulnerabilities before /login. The CLI needs a session before it can read your Safeguard data. /whoami confirms who you are signed in as.

Starting the CLI in the wrong folder. The @ picker and /resume both work from the folder you launched in. If a file is missing from the picker, quit and restart in the right directory.

Leaving Yolo on. Yolo approves every action, including commands. Switch back to Auto with /permissions when you are done with a scratch folder.

Doing heavy editing on the phone. The web app docs point mobile at review and alerts. Write policies and inspect SBOMs on the Mac.

FAQ

Is there one Safeguard app for both iPhone and iPad?

Yes. The download page lists the iOS app on the App Store for iPhone and iPad. Install it on each device and sign in with the same account.

Is the Safari extension available on Mac?

Not yet. The download page lists Safari on Mac as in Apple review. On iPhone and iPad the Safari extension is included in the app; turn it on in Settings, Apps, Safari, Extensions.

Which Mac do I need for the desktop app?

The Apple Silicon build needs macOS 12 or later. Intel Macs use the x64 build linked from the same download section. Both are on safeguard.sh/download.

Will macOS warn me when I open the desktop app?

The download page says the Mac build is signed with Safeguard.sh Inc's Developer ID and notarized by Apple, so there are no warnings to click through.

Can I keep my Claude, ChatGPT or Codex plan if I switch to Safeguard?

Yes. Bring the Claude, ChatGPT or Codex plan you already pay for and use it inside the Safeguard CLI. Safeguard adds no extra charge for that usage; your plan's own usage limits and terms still apply. You can also connect Claude or ChatGPT to the Safeguard MCP server and use Safeguard from the chat you already use. See the CLI docs or /model for connecting a model.

Does my source code leave the Mac?

When the local runner runs platform scan jobs, your source code stays on the Mac. Only SBOMs and findings are uploaded.

Can the CLI change files on my Mac without asking?

Writing files, editing code and running commands ask before they act, according to your /permissions mode. In Auto, the default, known actions are approved and anything new is asked. Only Yolo approves everything, and you choose it yourself.

What should I use on the iPad versus the Mac?

Use the iPad for review: checking findings, reading what changed and approving proposed remediations. Use the Mac for the work itself: the CLI to write and fix code, and the desktop app for heavier tasks such as policy editing and SBOM inspection.

Try it yourself

The Mac part of this guide starts with the CLI. This walkthrough shows how to set up the runner on your own machine, pick the platform, start it, approve it with its device code, then use the same CLI to scan from the terminal.

Click through it at your own pace, or open it full screen.

Ready to set up your own devices? Start at safeguard.sh/download for the Mac app and the App Store link, install the CLI with one line in Terminal, and keep the CLI docs open for your first session.

Quick check

Sources

• Safeguard downloads
• Safeguard CLI overview
• Safeguard CLI interactive terminal
• Safeguard local runner
• Safeguard desktop application
• Safeguard web application
• Safeguard MCP integrations
• Connect Claude to Safeguard MCP
• Connect ChatGPT to Safeguard MCP
• Claude download page
• Anthropic: custom connectors using remote MCP
• Apple: use Safari extensions
• Apple: multitask on iPad with iPadOS 26

Last reviewed: 21 October 2026.