AMOS144

AMOS144 / ZeroToken

Public

ZeroToken — Record once, automate forever. A lightweight MCP for agent-driven browser automation, script recording, and scheduled execution.

241
14
100% credibility
Found Mar 12, 2026 at 162 stars -- GitGems finds repos before they trend. Get early access to the next one.
Sign Up Free
AI Analysis
Python
AI Summary

ZeroToken enables AI agents to record browser actions once and replay them as efficient scripts, reducing costs and speeding up repetitive web automation tasks.

How It Works

1
🔍 Discover ZeroToken

You hear about a smart helper that lets AI assistants handle boring web tasks like logging in or checking reports, by recording once and repeating forever.

2
📦 Set it up quickly

You add the simple tool to your computer with one easy step, and get the web viewer ready.

3
▶️ Start the background helper

You launch a quiet service that keeps the web viewer alive and ready for your AI friend.

4
🔗 Link it to your AI assistant

You tell your AI tool where to find the helper, so they can team up smoothly.

5
🎥 Record your task once

Your AI performs the web steps like opening a site or clicking buttons, and everything gets saved as a perfect replay guide.

6
✨ Turn recording into magic repeat button

The saved steps become a simple button you can press anytime to redo the task without asking the AI to think again.

🚀 Automate forever, save time and money

Now your daily logins or data checks happen automatically, fast and cheap, every time you need them.

Sign up to see the full architecture

5 more

Sign Up Free

Star Growth

See how this repo grew from 162 to 241 stars Sign Up Free
Repurpose This Repo

Repurpose is a Pro feature

Generate ready-to-use prompts for X threads, LinkedIn posts, blog posts, YouTube scripts, and more -- with full repo context baked in.

Unlock Repurpose
AI-Generated Review

What is ZeroToken?

ZeroToken is a Python-based lightweight MCP server for agent-driven browser automation using Playwright. It lets AI agents record browser scripts once—capturing actions like clicks and inputs with screenshots and state—then replay them forever deterministically, slashing token costs on repetitive tasks. Run it via CLI as `zerotoken-mcp-http` for persistent HTTP sessions or stdio, integrating seamlessly with tools like OpenClaw.

Why is it gaining traction?

It stands out by turning ad-hoc agent browser interactions into stable, scheduled scripts via MCP tools like trajectory_start, trajectory_to_script, and run_script_by_job_id. Built-in smarts handle flaky sites with adaptive element relocation, smart waits, and error recovery, while stealth mode dodges anti-bot detection. Developers love the zero-LLM-execution mode for automation that just works across sessions.

Who should use this?

AI agent builders wiring browser tools into OpenClaw or Cursor IDEs for tasks like daily logins or data grabs. RPA scripters needing reliable web replay without constant LLM prompts. Backend devs automating scheduled browser execution, like report pulls or form submissions.

Verdict

Grab it if you're prototyping agent-driven browser flows—solid docs and MCP integration make it dead simple to start, despite 14 stars and 1.0% credibility signaling early maturity. Polish tests and add examples for production; it's a smart bet for token-thrifty automation now.

(198 words)

Sign up to read the full AI review Sign Up Free

Similar repos coming soon.