Recommendation
No need for programming, easily automate your computer work.
Core content:
1. autoMate: AI+RPA automation tool, free your hands
2. Built on OmniParser, supports natural language interaction
3. Deployment guide: GitHub download, environment construction and startup method
Yang Fangxian
Founder of 53AI/Most Valuable Expert of Tencent Cloud (TVP)
Free your hands! No programming knowledge is required, this open source tool can make your computer work for you.When I was browsing GitHub recently, I found an eye-catching project: autoMate. It has been open source for 3 weeks and has already received 3000+ Stars . I call it a magic tool for workers to fight repetitive work.Imagine this: before leaving get off work every day, you say to your computer "help me organize data, make reports, and reply to emails", and all the work will be completed automatically when you go to work the next day.01
Project Introduction
autoMate is a revolutionary AI+RPA automation tool built on OmniParser, which aims to enable computers to complete repetitive tasks instead of humans through natural language interaction.It combines AI big models and interface operation technology, can understand user instructions, make autonomous decisions and simulate human operation of computer interfaces, helping users free themselves from tedious work and focus on high-value matters. It can:? Understand your needs and automatically plan tasks? Intelligently understand screen content and simulate human vision and operation? Make decisions, judgements and take actions independently based on task requirements? Support local deployment to protect your data security and privacyThe developer created a demo: after deploying autoMate, select the WeChat client part and give instructions to the AI, it will automatically approve friend requests and tag friends.This case is very suitable for people who do private domain marketing. It is based on Microsoft's Omniparser + AI big model, which can automate many scenarios.Omniparser is an AI tool based on pure vision. Simply put, it is like a pair of intelligent eyes that can recognize and understand every interactive element on the computer interface , such as buttons, input boxes, icons, etc. through screenshots.Whether it is a Windows pop-up window, a web form, an Office document editor, or a mobile APP interface, it can see everything clearly. For more introduction, please refer to my previous articles.02
How to deploy
You can download the executable version directly from github release.Open source address: https://github.com/yuruotong1/autoMate/releases/tag/v3.0
If you cannot access this webpage, you can follow this public account and send a private message to 250413 to obtain the download address accessible in China.The developers strongly recommend installing miniConda first and using miniconda to install dependencies. There are many tutorials online, or you can ask AI for help. Then follow the following commands to set up the environment:
# Clone the project
git clone http s : //github.com /yuruotong1/autoMate.git
cd autoMate
# Create python3 . 12 environment
conda create -n "automate" python == 3.12
# Activate environment
conda activate automate
# Install dependencies
python install.py
Once installed, you can launch the application using the command line:
After successful execution, open the browser http://localhost:7888/ to configure the AI large model API key and basic settings. Currently, only the underlying models of Open AI are supported, such as o1 and GPT-4o.