Introducing ActionWeaver, a library for scaling and orchestrating OpenAI functions for LLM agent

Hey guys,

I’m excited to share a project I’ve been working on and would greatly appreciate your feedback!

ActionWeaver is a cool library that makes it easy to scale up and orchestrate OpenAI functions. With a simple decorator, you can turn ANY regular Python code into OpenAI functions. Moreover, you can go beyond the OpenAI function limit by constructing hierarchies and chains of functions, adding a significant boost to agent capabilities.

Here is a quick demo built with ActionWeaver, OpenAI functions, and vanilla Python code.

5 Likes

That looks super powerful, hopefully I’ll get some time to try it out.

1 Like

Thanks! Keep me posted on how it’s working and any ideas to make it solve problems even better.

Hey, I’m starting the process of migrating my scripts and personas into agents.
I’m excited and hoping for good results.

1 Like

Awesome! Let me know if you run into any issues, love to collaborate!

Would love to collaborate with you on this. You certainly will know the ends and outs.

1 Like

A RAG bot can retrieves content on demand, it uses ActionWeaver to orchestrate llama index and langchain tools to implement a new search experience.

Github

Interactive RAG Demo

2 Likes

Hello @Foxalabs, wondering if you had the opportunity to take a look at it. Love to hear your feedback.

1 Like

Not yet, It’s on my list! Hopefully have a project that I’ll try it out with coming up soon.

2 Likes

leaving a note. tocome back for this over the weekend.

ActionWeaver, AI application framework that put function-calling as first-class feature, has just launched a new version, supporting OpenAI API and Azure OpenAI service!

1 Like

@Foxalabs @brianjking We just released 0.0.16 that supports parallel function calling !!!

I’d like to invite you guys to give it a shot and let us know your opinions !

1 Like