Hacker-Valley-Media/Interceptor

Browser automation and computer use for AI agents. One CLI and MCP server for your signed-in browser, native macOS apps, and real iPhones.

What it solves

Interceptor provides a way for AI agents to interact with a user's actual operating environment. It solves the problem of agents being trapped in sandboxes or separate automated browser instances, allowing them to use the user's existing signed-in browser sessions, native macOS applications, and physical iPhones to perform real-world tasks.

How it works

Interceptor acts as a bridge between an AI agent (via CLI or a Model Context Protocol server) and the device's hardware and software. It uses a combination of browser extensions, accessibility controls, Apple Events, and a custom XCUITest runner for iOS to read UI structures (DOM, element trees) and execute actions (clicks, typing, gestures). It supports background execution for macOS and browser tasks so the user can continue working while the agent operates.

Who it’s for

It is designed for developers building AI agents that require deep integration with a user's desktop and mobile device, as well as users who want to automate complex workflows across web and native apps.

Highlights

  • Cross-Platform Automation: Supports browser automation (Chrome, Brave, Safari) on Windows, Linux, and macOS, with native macOS and physical iPhone control available on Mac.
  • Existing Sessions: Operates within the user's current browser profiles, cookies, and logins rather than launching a clean slate.
  • Deep Inspection: Goes beyond pixels to read DOM, accessibility trees, network traffic (XHR/WebSocket), and even native app internals.
  • Research Workflow: Includes a specialized research skill with query planning and an evidence ledger for browser-based investigations.
  • Record and Replay: Allows users to demonstrate a workflow once and export it as a reusable plan for the agent to follow.
  • MCP Integration: Built-in support for the Model Context Protocol (MCP) for easy integration with clients like Claude Desktop or Cursor.

Related

  • Project
    AgentiLoop/Agent

    AgentiLoop Agent is a native macOS app that lets you converse with any major LLM and have the model autonomously perform real‑world tasks—coding, GUI automation, shell commands, web browsing, voice control, and custom Swift scripts—through a self‑verifying task loop and a rich toolbox of built‑in tools.

  • Project
    To3akaRin/mac-computer-use

    mac-computer-use is an open‑source Agent Skill that lets LLM‑driven agents observe, control, and verify macOS UI elements (via Accessibility APIs or Chrome DevTools Protocol) from a local CLI, enabling autonomous desktop automation on macOS 14+.

  • Dispatch
    Safari Technology Preview 247: Introducing the Safari MCP Server

    Apple introduces the Safari MCP server in Safari Technology Preview 247, allowing AI agents to directly interact with the browser for automated debugging, performance analysis, and accessibility testing.

  • Project
    uvwt/agentdock

    AgentDock is an open‑source runtime that lets LLM chat agents (e.g., ChatGPT) safely perform file operations, run commands, use Git, automate browsers, and coordinate work across multiple computers or containers via a secure HTTP‑based MCP endpoint. It provides fine‑grained permission controls, task persistence, and extensible Skills, but does not include its own chat UI or model inference.

  • Project
    getsentry/MobileBuildMCP

    MobileBuildMCP is an open‑source tool that runs an MCP server and a CLI to let AI agents (or humans) invoke Xcode build and test commands for iOS/macOS projects. Install via Homebrew or npm, start the server with `mobilebuildmcp mcp`, and use subcommands like `simulator build` to compile or test apps. Optional “skills” JSON files teach LLM assistants how to talk to the server. MIT‑licensed, macOS‑only, with Sentry telemetry.