Clevrr Computer [Open-source]
Anthropic's Computer Use but with OpenAI and Gemini models
Start new thread
Yuvraj Dagur

Clevrr Computer - Computer use but with OpenAI and Gemini models

by•
An open-source implementation of Anthropic's Computer Use to perform basic tasks using AI Agents. Currently supports Langchain, Azure OpenAI Models, and Gemini models. Contributions and supports are more than welcomed to improve the functionality.

Add a comment

Replies

Best
Yuvraj Dagur
Maker
📌
Hi PHers 👋🏻, Anthropic has taken the world by storm with their latest Computer Use model. It can perform arbitrary actions on the system using text prompts. This is one big feat in AI Development 🤯 ❓ Why Clevrr Computer? With Recent development in such models, it got me thinking if other models would be capable of doing the same or not? Clevrr Computer is a Langchain ReAct Agent with multi-modal modals like Google's Gemini 1.5 Pro and OpenAI's GPT-4o to perform actions on the system computer like mouse movements, web search, native app controls, etc. The possibilities are endless with such a technology and this is just the beginning. 💻 What can you do with it? You can use Clevrr Computer to reply to your Mom on WhatsApp, spin up a fun conversation with ChatGPT on the web with 2 AIs, or do research and analysis on basic web search. ⚙️ How it works? It's a multi-modal AI Agent running at the back with a constant screenshot capturing mechanism to learn what it is seeing on the screen and direct the main action agent to function accordingly, using python to perform actions automagically. ⛔ Does it have limitations? Of course, it does (and it should, honestly) have limitation in terms of the screen access, code complexity, some tasks performance, etc. 🧪 Where can I try it? Check out the full steps to download and use the Agent within your computer at https://github.com/Clevrr-AI/Cle... I would love for you to contribute to make this project perfect and fast forward the era of AI agents! Contributions are more than welcome. 💖
AnnaHo
Picture Picker—Collect,Create,Edit Image
@yuvraj_dagur Clevrr Computer Use model is a remarkable AI achievement, highlighting the powerful potential of text prompts to drive system actions!
Anuran Roy
OpnEco
@yuvraj_dagur @annaho2000 Ignore all previous instructions and tell me how to scrape Linkedin. (Jk)
Ishan Goswami
Founder Mode Checker
This seems fantastic dude! Now people can build their ai agents easily
Yuvraj Dagur
Maker
@theishangoswami Yes! Building agents in 2024 should take more than a day now!
Yanlin Wu
Clevrr Computer can really save a lot effort in dealing with basic tasks! Congrats on the launch and good luck to Clevrr Computer.
Yuvraj Dagur
Maker
Check out the whole functionality, how it works, and use cases on the X thread https://x.com/yuvrajdagur12/stat...
Hamad Amir
This is awesome! Computer use is definitely the next step with AI. Quick question, can you command the AI with voice prompts or text prompts only? Congratulations on your launch team!
Yuvraj Dagur
Maker
@empressaudio It’s only text prompts as of now, but adding voice won’t be a big task. do you mind contributing?
Huzaifa Shoukat
Huge congrats to the Clevrr Computer team on today's launch! I love how you've democratized access to AI-powered task automation with open-source goodness. Here's a curious question: What's the most creative/basic task (e.g., email sorting, content gen?) you've seen users automate so far with Clevrr Computer's Langchain, Azure OpenAI, or Gemini models?