Overview
What browserlane is, why it drives Chrome over WebDriver BiDi, and who it is for.
browserlane is browser automation for humans and AI agents. It is one Rust
binary, bl, that exposes two surfaces over the same engine:
- A CLI for humans and scripts —
bl go,bl click,bl screenshot,bl find, and dozens more. - An MCP server for AI agents —
bl mcpspeaks JSON-RPC 2.0 over stdio, so an agent can drive a real browser through a catalog of tools.
Both surfaces drive Chrome over the WebDriver BiDi protocol.
Why WebDriver BiDi
Most agent-facing browser tools sit on the Chrome DevTools Protocol (CDP) — Chrome's vendor-specific side-channel, the same plumbing Playwright and Puppeteer use underneath. browserlane instead drives Chrome exclusively over WebDriver BiDi, the bidirectional, W3C-standard protocol the browser vendors are converging toward. Building on the standard rather than a vendor side-channel is the core bet.
What makes it different
One static binary, no runtime
bl ships as a single native executable for macOS, Linux, and Windows.
There is no Node or JVM runtime to install — download it, put it on your
PATH, and run.
Visible browser by default
Commands open a real, visible Chrome window so you can watch automation
happen. Pass --headless when you want it hidden.
Auto-installs Chrome
On first use, bl downloads a pinned build of Chrome for Testing into a
local cache. Run bl install to fetch it ahead of time.
Two surfaces, one engine
The human CLI and the agent MCP server share the same browser engine, so an automation you prototype by hand behaves the same when an agent runs it.
Two quickstarts
Pick the on-ramp that matches how you want to use browserlane:
Drive a browser yourself (CLI)
Run real commands — navigate, screenshot, map elements, click, read text, and evaluate JavaScript.
Wire up an AI agent (MCP)
Register the MCP server with your coding agent and let it drive Chrome with natural-language prompts.
New to browserlane? Start with Install.