Skip to main content
BrowserWire is also available as a native desktop application built on Electron. The desktop app gives you a fully integrated experience: an embedded Chromium browser pane, discovery controls, session history, and action testing — all in one window, with no need to separately install the Chrome extension.

​
What’s different about the desktop app

With the CLI, you install the Chrome extension separately and use a Chrome sidepanel to control discovery. The desktop app replaces both:
  • The embedded browser pane is fully instrumented — no extension installation needed
  • Discovery controls sit alongside the browser in the same window
  • The same REST API runs on port 8787, so any agent or tooling that works with the CLI also works with the desktop app
The desktop app and CLI share the same manifest store, so sites discovered in one are immediately available in the other.

​
Getting started

Download the latest release for your platform from the BrowserWire releases page. Open the downloaded file and follow the standard installation steps for your OS.
On macOS, if you see a Gatekeeper warning on first launch, right-click the app and choose Open to bypass it.

​
Configuring your LLM provider

Before you can run a discovery session, you need to configure an LLM provider. The desktop app needs a vision-capable model to analyze page content during discovery.
  1. Open Settings from the menu bar (BrowserWire → Settings) or press Cmd+, / Ctrl+,
  2. In the Settings panel, select your LLM provider (OpenAI, Anthropic, Gemini, or Ollama)
  3. Enter your API key
  4. Save your settings
ProviderDefault modelRequires API key
OpenAIgpt-4oYes
Anthropicclaude-sonnet-4-20250514Yes
Geminigemini-2.5-flashYes
Ollamallama3No
Ollama lets you run a local model with no API key. It’s a good option for offline use or if you want to avoid sending page content to a third-party provider.

​
Panels

The desktop app has four main panels, accessible from the activity bar on the left side of the window.

​
Discovery

The default view. The embedded browser takes up most of the window, with a control panel on the right.
  • Use the address bar at the top to navigate to a site
  • Click Start Exploring to begin a session — the browser is now instrumented and capturing
  • Browse the site normally: navigate pages, open menus, interact with forms
  • Click Stop Exploring when done — the app processes the session and builds the manifest
The right panel shows a live snapshot count while exploring, so you can see that captures are happening.

​
History

Review past exploration sessions. Select a session to see what was captured: which pages were visited, how many snapshots were taken, and the resulting manifest content. After you stop exploring, the app switches to the History panel automatically and selects the session you just completed.

​
Execution

Test discovered actions and workflows directly from the desktop app, without needing to use curl or a separate API client. Select a site, choose an operation, fill in any inputs, and run it — the embedded browser executes the action on the live page.

​
Settings

Configure your LLM provider, API key, and other preferences. Changes take effect for the next discovery session.

​
The REST API

The BrowserWire server runs on port 8787 while the desktop app is open, exposing the same REST API as the CLI: 
# List discovered sites
curl http://localhost:8787/api/sites

# Get the OpenAPI spec
curl http://localhost:8787/api/sites/example-com/openapi.json

# View interactive Swagger docs
open http://localhost:8787/api/sites/example-com/docs
Any agent or integration built against the CLI API works identically with the desktop app.