R1 as an 'Agent'

Hey there! A few days ago, I came across Lindy.ai and was really impressed by how simple it is to create complicated automations and routines for agents.

It’s incredibly powerful to set these up and have them handle ongoing, complex tasks. The node-based interface makes designing these flows straightforward, though I think it’s particularly appealing to power users. It would be amazing if we could just describe what we want, and have an agent understand, improve and execute it.

I’d love for my personal assistant to take over multi step task, working on my behalf in the background. I’m really looking forward to seeing some updates in R1 that will make it possible to create or use a more agentive rabbit assistant.

4 Likes

thanks for sharing this, i ll check it out

1 Like

So many cool integrations :slight_smile:

2 Likes

I am keeping an observation attitude on this ai at the moment, and I need to check the background of this ai. :thinking:

1 Like

You mean Lindy? You can try it out.

1 Like

Lindy.ai, that’s really good.

Thanks for sharing the lindy.ai link @Designer ! AutoGPT released a similar graphical / flow based setup, albeit with fewer integrations to choose from and a bit more technical/low level (but open source, no subscription).

A goal based agent like @Designer envisions for the R1 is implemented in SuperAGI (yes, ambitious) in which you configure the tools that are available for the agent (eg email access, web scraping, Jira access) which then are used by the agent as needed to work towards the goal given by the user. Drawback: it needs many ChatGPT API calls to reach the goal making it expensive in use.

1 Like