Interceptor CLI
A CLI and MCP server that give an agent your signed-in browser, desktop apps and a physical iPhone
- 577 ★ on the repository
- Last commit
Broad access
Why this level
- Commands run in your accounts and apps with saved cookies and logins, and the extension's permissions are broader than a single task
- Reads network traffic, cookies, storage and screens, can deliver a secret from the macOS vault and run code on the page and on the phone
- On macOS it installs a daemon and a native bridge that are granted Accessibility, screen recording and microphone permissions
Install
- Works in
In your terminal, with SkillFoxx CLI
npx skillfoxx add cli/interceptorDetects the agents on your machine, shows the access level and pins the version.
Other ways to install
Access and checks
How we reviewBroad accessThe tool can change data, run commands or handle secrets.
- Commands run in your accounts and apps with saved cookies and logins, and the extension's permissions are broader than a single task
- Reads network traffic, cookies, storage and screens, can deliver a secret from the macOS vault and run code on the page and on the phone
- On macOS it installs a daemon and a native bridge that are granted Accessibility, screen recording and microphone permissions
This is third-party code. Review the repository files before installing.
What it does
Interceptor connects an agent or script to your signed-in browser, native apps on macOS, Windows and Linux, and a physical iPhone. It is the `interceptor` CLI with a local daemon and an extension for Chrome, Brave and Safari. No separate automated browser is launched: the agent works in your own profiles. `interceptor open "https://example.com"` opens a tab in the background and returns the page tree and text. `interceptor act <ref>` clicks an element by the reference from read or find, or types a text argument into it. The agent reads the DOM, tables and forms, captures network traffic (fetch, XHR, SSE, WebSocket) and takes screenshots with OCR. Editors like Google Docs, Sheets and Canva get dedicated interfaces, and the `--group` flag keeps the background tabs of parallel tasks apart. The Full package adds a macOS bridge. Through it the agent reads accessibility trees, clicks and types in windows and menus, and works with PDF, calendar and contacts. For the iPhone Interceptor installs its own XCUITest runner, which shows the element tree, performs taps and gestures and takes screenshots. `interceptor monitor start` records your workflow, and `interceptor monitor export <session-id> --plan` outputs a replay plan. Interceptor reaches an agent two ways. `interceptor skills adopt` installs the bundled skills, and `interceptor mcp install` registers the MCP server in detected clients.
Who it is for. For developers and researchers who want to let Claude Code, Codex or Cursor into their signed-in browser, native Mac apps and a real iPhone.
Good fit when
- You need browser research with your logins: page text, tables, links, search and a source ledger in the research skill
- An agent must edit a document, slides or a design in a web editor and save the result, for example an export file from the page
- The task is in a native Mac app: Finder, windows, menus and dialogs in the background while you keep working
- You need to test an app on a real iPhone: element tree, taps, input, screenshots and process load
- You demonstrate a workflow once, record it with `interceptor monitor start` and get a replay plan
Not a fit when
- You need an isolated headless browser for CI: Interceptor works through your browser with an extension and a local daemon
- You are not ready to let an agent into your accounts and personal browser data
- You have an Intel Mac and cannot build from source: the ready macOS packages target Apple silicon
- You need an open-source license: the Elastic License 2.0 is not one
Example request
Open my signed-in browser session, read the pricing table on the vendor page and write the prices as Markdown, without clicking or submitting anythingLimitations
- Both components are needed: the local runtime and the browser extension. Safari takes a separate package, and you enable the extension in Safari settings yourself.
- Ready macOS packages target Apple silicon, and the native bridge needs macOS 15 or newer. The Windows installers target Windows 11 24H2 and newer.
- The iPhone needs a Mac with the Full package, a phone in Developer Mode, Xcode with an Apple ID and an unlocked screen during work. There is no background operation as on the computer.
- Standard reads mask password fields, but screenshots, scene dumps, `eval` and network data are not redacted automatically. The agent may send the content it receives to its model provider.
- The project ships under the Elastic License 2.0, and commercial terms are described in COMMERCIAL.md.
How to disable. Disable the extension on the browser's extensions page and remove the MCP server from client configs. On macOS the package is removed with `sudo bash "/Library/Application Support/Interceptor/uninstall.sh"`, and the `--bridge-only` flag removes only the native bridge and keeps the browser part. On Windows use the standard program removal.
FAQ
How do I connect Interceptor to Claude Code or Codex?
After installing the package run `interceptor mcp install`: the installer detects Claude Code, Codex, Gemini CLI, Cursor and Claude Desktop and registers the MCP server in them. Then restart the client and check `interceptor mcp status`. For an agent that can run shell commands, `interceptor skills adopt` is enough.
How is the Full package different from Browser?
Browser gives the CLI, daemon and extension files for browser tasks. Full adds the native bridge: Mac app control and iPhone tooling, and on Windows and Linux the `interceptor desktop` command for native apps. Browser is enough for web tasks.
How do I check that Interceptor works in the browser?
Run `interceptor open "https://example.com" --group first-task`: you get the page tree and text with the Example Domain heading. Then `interceptor read --group first-task` reads the page again, and `interceptor group close first-task` closes that task's tabs. Extension diagnostics run with `interceptor diagnose extension`.
What permissions does the Interceptor MCP server get?
Reads and UI mutations are on by default. Tiers with destructive actions and arbitrary code execution need the operator's opt-in and ask for confirmation. These limits belong to MCP, and the direct CLI follows its own command contracts.
How do I control an iPhone through Interceptor?
You need a Mac with the Full package, a physical iPhone in Developer Mode, Xcode signed into an Apple ID and USB for the first pairing. `interceptor ios setup` builds, signs and installs the runner, then `interceptor ios tree` shows the element tree. A free Apple signature lasts 7 days, and Interceptor renews it.
Discussion
No questions or reviews yet.
Related
CLI toolsA CLI and MCP server that lets an agent drive your real Chrome with its logged-in sessions: 103 commands across 36 platforms plus browser automation
A remote MCP server and Chrome extension: the agent reads data behind your logins and operates the sites you already use
An MCP bridge to control your browser from a coding agent: a Chrome extension and a set of tab and page automation tools
A separate browser for agents with a skill and MCP: Claude Code and Codex work in your logged-in accounts, and every session is recorded for replay
An MCP server that drives your already open Chrome and stops to ask you about a 2FA code, a CAPTCHA, or a choice only you can make
A local browser automation platform: MCP server, CLI and skills so an agent can drive Chrome, fill forms and pull data from pages
Comparison with similar tools
Entries that share the main search query, and how they differ.
| Entry | Official | Authentication | Repository stars | Main difference |
|---|---|---|---|---|
| Interceptor (this entry) | yes | 577 | ||
| Browser MCP | yes | not required | 7,246 | an MCP server and Chrome extension that let an agent drive an already open browser |
| bb-browser | no | 6,255 | a CLI and MCP server for your real Chrome with saved logins |
README in short
The README opens with a table of typical tasks and packages for macOS, Windows and Linux, then covers installation, the first browser task and connecting an agent. Separate sections cover browser automation, macOS, iPhone, workflow recording and deeper interfaces like Electron and the native runtime. The end collects permissions and data, updates, diagnostics and building from source.