What is New Order Global?

New Order Global is the lightest autonomous browser agent at around 1 MB. It is an AI-powered Chrome extension that lets you create any custom browser tool and automate any task using plain English. It includes a built-in free YouTube customization tool, an AI tool builder, and the Global Executive — an autonomous browser agent with 40+ supported actions including web search, PDF handling, email reading, file capture, and more. The source code is available on GitHub for self-hosting.

Is New Order Global the fastest browser agent?

Yes. New Order Global is the fastest browser agent available. At around 1 MB, it is the lightest autonomous browser agent — built with vanilla JavaScript and Chrome Manifest V3 with zero build step. It loads instantly, injects into pages in milliseconds, and executes 40+ actions with minimal overhead. Unlike heavy frameworks, it runs directly in your browser with no external dependencies.

Is New Order Global open source?

New Order Global is source-available — the complete source code for both the Chrome extension and the server is available on GitHub (github.com/bckflpboys/new-order-global and github.com/bckflpboys/global-order-server-opensource). You can download, inspect, self-host, and run your own instance. While the license is not a traditional open-source license, the full source code is publicly accessible.

How much does New Order Global cost? Is it the cheapest browser agent?

New Order Global is the cheapest browser agent available. Every new account gets 10 free credits. One-time credit packages start at just $4 for 40 credits. Subscriptions start at $6.99/month for 150 credits. Credits are based on actual AI token usage, so lightweight models like Gemini Flash cost as little as 0.08 credits per tool — meaning 40 credits can create 500+ tools. The free tier includes the full YouTube New Order tool, 15 agent steps per task, and 10 daily tasks.

How is New Order Global different from other browser agents?

New Order Global is the only browser agent that runs as a native Chrome extension (~1 MB). Other browser agents like browser-use or Playwright-based agents require Python environments, heavy dependencies, and run outside the browser. New Order Global runs inside your browser with 40+ actions, multi-model AI support (Gemini, Claude, GPT-4o), Telegram/WhatsApp integrations, scheduled tasks, long-term memory, PDF handling, multi-agent council mode, and sub-agent spawning. It is source-available and self-hostable.

What tasks can New Order Global automate?

New Order Global can create any tool and automate any task. Examples include: web scraping and data extraction, form filling and submission, research across multiple websites with source verification, price monitoring, lead generation, document creation, message automation via WhatsApp and Telegram, PDF reading and editing, file downloading, email reading for OTPs, scheduled recurring tasks, and any multi-step browser workflow you can describe in plain English.

What AI models does New Order Global support?

New Order Global supports multiple AI models through OpenRouter: Google Gemini 2.5 Flash, Anthropic Claude, OpenAI GPT-4o, and more. You can switch models per task or per conversation. The Super Agent tier even allows assigning different models to different council roles (Strategist, Executor, Critic, Optimizer) for enhanced task planning.

Can I self-host New Order Global?

Yes. Both the Chrome extension source code and the backend server are available on GitHub. The server runs on Node.js with Express and MongoDB. You can self-host with your own OpenRouter API key, eliminating any dependency on the hosted service. This gives you complete data sovereignty and no rate limits.

Does New Order Global work with Telegram and WhatsApp?

Yes. New Order Global integrates with both Telegram and WhatsApp. You can send tasks to your browser agent via Telegram DM or a dedicated WhatsApp group, receive real-time progress updates, and control the agent with slash commands (/help, /cancel, /tasks, /copilot, /autopilot). The agent can also run tasks in the background without the browser panel being open.

Was New Order Global built with VoxCode?

Yes. Parts of New Order Global were built using VoxCode.app — the first mobile IDE and voice IDE. VoxCode allows developers to write and edit code using voice commands on mobile devices, making it possible to code from anywhere. Visit voxcode.app to learn more.

How does New Order Global create custom Chrome extensions?

New Order Global uses AI to generate Chrome extension code from plain English descriptions. You describe the tool you want (e.g., 'Build a price tracker for Amazon'), choose an AI model (Gemini, Claude, or GPT-4o), and the AI generates the JavaScript, CSS, and configuration. You preview it, iterate with feedback, then accept — and it auto-injects on matching websites.

How does the autonomous browser agent work?

The Global Executive browser agent works in a loop: you describe a task in plain English, the AI classifies it (research, action, or mixed), creates a plan with milestones, then executes 40+ actions autonomously across multiple browser tabs. Actions include page reading, clicking, typing, web search, PDF handling, email reading, file capture, and more. The agent reports each step back to the AI for the next action until the task is complete.

Do I need to know how to code to use New Order Global?

No. New Order Global is designed for non-technical users. You describe everything in plain English — the AI handles all the coding. The extension is also only ~1 MB and requires zero build steps to install.

Can I control the browser agent from my phone?

Yes. New Order Global integrates with Telegram and WhatsApp. You can send tasks, receive progress updates, and control the agent with slash commands from your phone. Tasks run in the background on your desktop browser even when the panel is closed.

How It Works

From Plain English to Autonomous Execution

New Order Global is the lightest autonomous browser agent at ~1 MB. Describe what you want in plain English — AI creates any tool and automates any task with 40+ browser actions. No coding, no build steps, no heavy dependencies.

Start Building Free View Documentation

Two Ways to Automate

Create permanent browser tools or run one-off autonomous tasks — both powered by AI, both described in plain English.

Path 1

AI Tool Builder

Create permanent custom Chrome extensions that auto-run on specific websites. Build price trackers, data collectors, UI customizers — any tool you can describe.

Describe

Tell the AI what tool you want. Choose Gemini, Claude, or GPT-4o.

Preview

AI generates JS + CSS + configuration. Test and iterate with feedback.

Deploy

Accept the tool. It auto-injects on matching websites with cloud sync.

Path 2

Global Executive — Browser Agent

Autonomous multi-step task execution across browser tabs. The fastest browser agent with 40+ actions — research, scrape, fill forms, send messages, handle PDFs.

Describe the Task

'Research the top 5 budget laptops under $800 from 3 different sites.'

AI Plans & Executes

Classifies as research/action/mixed, creates milestones, executes autonomously.

Get Results

Agent delivers structured results with verified sources. Review in Action Replay.

The Agent Execution Loop

The fastest browser agent uses a perception-planning-execution loop with 40+ actions. Here's exactly how a task runs from start to finish.

Step 1

Task Input

You describe the task in plain English. The agent receives the active tab's URL and content as context.

Step 2

AI Planning

The LLM classifies the task (research/action/mixed), sets milestones in the goal ledger, and returns the first action as structured JSON.

Step 3

Action Execution

The agent-runtime.js executes actions in the browser: click, type, scroll, extract, search, capture files, fill PDFs, read emails.

Step 4

Loop & Complete

Results feed back to the LLM for the next action. Research gate ensures thorough sourcing. Loop continues until done or step limit reached.

What Makes It the Fastest Browser Agent

Built as a native Chrome extension — no Python, no Playwright, no heavy frameworks. Just ~1 MB of pure browser-native performance.

~1 MB Footprint

The lightest autonomous browser agent. Vanilla JavaScript, Chrome Manifest V3, zero build step. Loads instantly and runs with minimal overhead.

40+ Browser Actions

Page reading, clicking, typing, scrolling, web search, PDF handling, email reading, file capture, coordinate-based actions, sub-agent spawning, and more.

Multi-Model AI

Switch between Gemini 2.5 Flash, Claude, GPT-4o, and more via OpenRouter. The Super Agent tier assigns different models to council roles.

Research Verification

Server-side web search with claim verification. Each research note is checked against source URLs — verified ✓, partial ⚠, or not found ✗.

Long-Term Memory

The agent remembers facts about you across tasks. Auto-extracted or manually added, with domain scoping and secret filtering.

Telegram & WhatsApp

Control the agent from your phone. Send tasks, receive updates, use slash commands. Tasks run in the background without the panel open.

Scheduled Tasks

Set up recurring browser automations with 5-field cron expressions. Runs unattended with pre-supplied briefing and autopilot mode.

Source-Available

Full source code on GitHub. Self-host both the extension and server. Your data, your infrastructure, your rules.