OpenCLI

// Overview

OpenCLI gives one surface to three kinds of automation. Built-in adapters expose sites such as Hacker News, Reddit, Twitter/X, Bilibili, Zhihu and Xiaohongshu as ordinary commands — `opencli hackernews top --limit 5` — so a scripted task gets a deterministic interface instead of a live browser session. Where no adapter exists, `opencli browser` primitives let an agent navigate, click, type, extract and inspect any page directly.

The key design choice is that it drives the Chrome you are already signed into. A lightweight Browser Bridge extension plus a small local daemon connect OpenCLI to your profile, so an agent works with your existing logins rather than re-authenticating or handling credentials. Multiple Chrome profiles are supported: you list the connected contexts, give them aliases, and select one per command — with several connected and no default set, OpenCLI asks rather than guessing.

It is also a hub for things that are not websites. Local binaries can be registered and exposed through the same discovery surface (`gh`, `docker`, Notion, Telegram and others), and desktop adapters cover Electron apps like Cursor, Codex, ChatGPT, Trae and Antigravity. For agents, the project ships installable skills — `opencli-browser` for ad-hoc page driving, `opencli-adapter-author` for writing a reusable adapter end to end, `opencli-autofix` for repairing a broken one, plus sitemap skills for navigating a site with recorded structure.

// What it does

• Built-in adapters turning popular sites into deterministic CLI commands
• `opencli browser` primitives — navigate, click, type/fill, extract, inspect — over your logged-in Chrome
• Browser Bridge extension plus a local daemon that auto-starts when needed
• Named Chrome profiles with per-command selection, so multi-profile setups are explicit
• Agent skills: opencli-browser, opencli-adapter-author, opencli-autofix and the sitemap pair
• CLI hub for local binaries (`opencli external register`) and adapters for Electron desktop apps
• Plugin system for publishing or installing third-party commands from a Git repo

// Getting started

Install the CLI, add the browser bridge extension, verify the setup with the built-in doctor, then run your first command. Node.js >= 20.18.1 is required for the npm install path.

node --version
npm install -g @jackwener/opencli

https://chromewebstore.google.com/detail/opencli/ildkmabpimmkaediidaifkhjpohdnifk

opencli doctor

opencli profile list
opencli profile rename <contextId> work
opencli profile use work
opencli --profile work browser main state

opencli list
opencli hackernews top --limit 5
opencli bilibili hot --limit 5

npx skills add jackwener/opencli

# or only what you need
npx skills add jackwener/opencli --skill opencli-browser

Commands and code are distilled from the project's own documentation — always check the official repo for the latest.

// When to use it

• Let an agent act on a site that has no API, using your existing session instead of credentials
• Replace brittle scraping with a deterministic command for a site you query repeatedly
• Write and verify a reusable adapter for a new site with the adapter-author skill
• Expose local binaries and Electron apps through one discovery surface an agent can enumerate

// How OpenCLI compares

OpenCLI alongside other open-source computer & browser use tools AI/TLDR tracks, ranked by GitHub stars.