Neura Market
GitHub logo

GitHub

Free

Automate your browser with GPT-4

FreeFree tier
Type
Open Source
Company
TaxyAI
Visit Website GitHub (40 stars)

About GitHub

Taxy AI is an open-source browser extension that leverages GPT-4 to automate repetitive browser tasks. It operates by extracting and simplifying the DOM of a webpage, identifying interactive elements, and sending a token-reduced version along with the user's natural language instruction to GPT-4. The AI then decides which actions (e.g., click, type, navigate) to perform. Currently in research preview, the extension supports ad-hoc instructions and runs entirely locally, ensuring no page contents or instructions are sent to external servers. It provides both a popup interface and a devtools panel for easy interaction. Users must provide their own OpenAI API key. Future plans include saved and scheduled workflows and a Chrome Web Store release.

Key Features

Uses GPT-4 to control and automate browser actions
Accepts ad-hoc natural language instructions
Extracts and simplifies the DOM to identify interactive elements
Templatizes the DOM to reduce token count
Runs locally with no data sent to external servers
Provides both a popup (cmd+shift+y / ctrl+shift+y) and a devtools panel interface
Open-source and self-hosted with your own OpenAI API key
Currently in research preview with more features planned

Pros & Cons

Pros
  • Completely open-source and transparent
  • Privacy-focused: no page contents or instructions sent to external servers
  • Free to use (only need your own OpenAI API key)
  • Can handle complex multi-step instructions
  • Easy to install locally with Node.js
Cons
  • In research preview; many workflows fail or confuse the agent
  • Requires an OpenAI API key (costs associated)
  • Currently only available via local build, not in Chrome Web Store
  • Installation requires Node.js >=16 and manual build steps
  • Limited to ad-hoc instructions; saved/scheduled workflows not yet available

Best For

Scheduling calendar events (e.g., 'Schedule standup tomorrow at 10am. Invite david@taxy.ai')Searching for and playing movies on NetflixProtecting branches on GitHubWriting an essay in the OpenAI playgroundGeneral repetitive browser task automation

FAQ

How does Taxy work?
Taxy pulls the entire DOM of a webpage, simplifies it to interactive elements, assigns IDs, and templatizes it to reduce token count. It sends the simplified DOM along with the user's instruction to GPT-4 to decide which actions to perform.
Is Taxy free to use?
Yes, the extension itself is free and open-source. You need to provide your own OpenAI API key, which may incur costs from OpenAI.
Can I install Taxy from the Chrome Web Store?
Not yet. Currently it is only available through the GitHub repository. You must build and install it locally. A Chrome Web Store release is planned after adding more usability features.
What are the system requirements?
You need Node.js >=16 and a Chromium-based browser (like Chrome).
Where is my OpenAI API key stored?
The API key is stored securely in your browser's local storage and is not uploaded to any third party.