I was wondering why cant I stop the agent when he does something wrong and not really following mij instructions it keeps in a loop that it wants to do it even when I try to stop it. Is there a way to talk to the agent even if its doing a task to steer it in the right direction.

📢 There will be an Agent Zero token, along with an official platform. Custom-built from scratch for shareholding and decentralized governance of the project. ❌ No memecoin, no pumpfun, no presale, no airdrop. ☝️Until the official detailed announcement, stay safe and avoid all the scam coins and fake profiles impersonating the project.

Have you thought about adding perplexity api sonar or tavily api

Hello, friends. Today I have a favor to ask. Seeing companies like OpenAI catching up to opensource projects with limited success makes me wonder how far ahead would have we been if we had more attention, people, funding. I am humbly asking for spreading word about A0 on social media. Especially now that people seem to see AI control something for the first time… they deserve to know. Thank you ❤️ And here's a post you can use to save you some time: For anyone thinking OpenAI Operator is something new, Agent Zero does that and much more. Opensource will always be ahead. Agent Zero 🤖 Can now use BROWSER 🌐 https://youtu.be/quv145buW74

I am pushing deepseek-reasoner support to testing branch. You just need to remove the temperature setting from model params in the settings as it does not support temperatures.