Skip to main content
This guide walks you through installing BrowserWire, connecting the Chrome extension, and making your first API call against a discovered site.
1

Install the CLI

Install BrowserWire globally with npm:
npm install -g browserwire
Or run it directly without installing:
npx browserwire
2

Load the Chrome extension

The Chrome extension is bundled with the CLI. First, get the path to the extension directory:
npx browserwire --extension-path
Then load it in Chrome:
  1. Go to chrome://extensions
  2. Enable Developer mode using the toggle in the top right
  3. Click Load unpacked
  4. Select the path printed by the command above
The Chrome Web Store release is in progress. For now, the extension must be loaded as an unpacked extension in developer mode.
3

Configure your LLM provider

BrowserWire uses a vision-capable LLM to perceive page content during discovery. Set your provider and API key as environment variables:
export BROWSERWIRE_LLM_PROVIDER=openai   # openai | anthropic | gemini | ollama
export BROWSERWIRE_LLM_API_KEY=sk-...
Supported providers and their default models:
ProviderDefault modelRequires API key
openaigpt-4oYes
anthropicclaude-sonnet-4-20250514Yes
geminigemini-2.5-flashYes
ollamallama3No
You can also set these in a .env file in your working directory, or in ~/.browserwire/config.json. See Installation for all configuration options.
4

Start the server

browserwire
The server starts on http://127.0.0.1:8787 by default. You can open the API docs landing page at http://127.0.0.1:8787/api/docs.
5

Discover a site

With the server running and the Chrome extension loaded:
  1. Open any website in Chrome
  2. Click the BrowserWire extension icon to open the sidepanel
  3. Click Start Exploring and navigate around the site
  4. Click Stop Exploring when you’re done — BrowserWire builds the manifest in the background
The site will appear in the API once discovery finishes.
6

Query the discovered API

List all sites you’ve discovered:
curl http://localhost:8787/api/sites
Fetch the OpenAPI spec for a specific site (replace example-com with your site’s slug):
curl http://localhost:8787/api/sites/example-com/openapi.json
The slug is the domain with dots replaced by hyphens — for example, example.com becomes example-com.
Point your tool-using LLM directly at the OpenAPI spec URL. It contains all discovered actions with typed input schemas your agent can call without any manual wiring.

​
What’s next

Once a site is discovered, you can execute actions through the REST API. For example, to call a submit_login workflow: 
curl -X POST http://localhost:8787/api/sites/example-com/workflows/submit_login \
  -H "Content-Type: application/json" \
  -d '{"email": "user@example.com", "password": "secret"}'
BrowserWire routes the action through the Chrome extension, which executes it on the live page using the discovered locators. You can also browse interactive Swagger docs for any discovered site at http://localhost:8787/api/sites/example-com/docs.