https://github.com/d3n7/GPT-4-Unlimited-Tools

Custom Plugins for GPT-4 (easy web app interface)
https://github.com/d3n7/GPT-4-Unlimited-Tools

Last synced: 3 months ago
JSON representation

Custom Plugins for GPT-4 (easy web app interface)

• Host: GitHub
• URL: https://github.com/d3n7/GPT-4-Unlimited-Tools
• Owner: d3n7
• Created: 2023-04-07T21:56:49.000Z (over 1 year ago)
• Default Branch: main
• Last Pushed: 2023-04-23T10:56:19.000Z (over 1 year ago)
• Last Synced: 2024-06-23T11:36:39.771Z (5 months ago)
• Language: Python
• Size: 78.1 KB
• Stars: 279
• Watchers: 11
• Forks: 37
• Open Issues: 0
• Metadata Files:
  • Readme: README.md

Awesome Lists containing this project

• awesome-gpt4 - GPT-4 Unlimited Tools - GPT-4 with internet and command line access. (Tools / Open-source projects)

README

        
# GPT-4 Unlimited

Custom Plugins for GPT-4 or GPT-3.5-Turbo (easy web app interface)

Current features:

- Googling information

- Computer Vision

- File read/write abilities

- Python script execution

Basically, this app gives GPT-4 access to your command line. (Don't worry! By default you must approve every command manually)

This unlocks the ability for GPT-4 to interact with the internet, APIs, python scripts, and any applications that you could with a CLI. Basically it's open source, flexible, plugins for GPT-4.

Web app: https://huggingface.co/spaces/cbg342/GPT4-Unlimited-Plugins (Windows users are advised to use the HuggingFace)

# Examples

Using BLIP-2 (via Replicate API) to give GPT-4 vision!



This information is accurate thanks to it's ability to Google:



Automatic file generation:







Trying to make it self aware lol



With a ***fully editable knowledgebase*** to tell GPT-4 what commands it has access to, how they are used, and how they should be run by your CLI.



# Installation And Startup

1. ```pip install -r requirements.txt```

2. If using vision plugin, add your Replicate API token to plugins/blip2.py

3. ```streamlit run app.py```

# To Do:

- Better parsing of incorrect command syntax instead of telling GPT to correct itself. (Save tokens)

- Rolling content window to avoid max tokens error