The video demonstrates the process of creating and deploying an AI agent using a tool called Archon. Initially, a request is sent to create a thread ID, which is crucial for subsequent interactions with Archon. This thread ID is then passed to Archon for all future calls. The system utilizes a custom API endpoint created for Archon. The demonstration shows the first API request being made, with a tool named 'wind surf' waiting for the response. The process involves a multi-step workflow within Archon: first, a Reasoner model establishes an initial scope for the agent. Subsequently, a coder agent is employed. This coder agent performs Retrieval Augmented Generation (RAG) to access and integrate information from the 'pantic AI documentation'. Finally, this combined information is used to construct the initial version of the AI agent, which is then successfully built using the code generated by Archon.

Cole Medin

So those are our two tools. Let's go ahead and send in this request.

So the first thing it'll do is call this tool to create our thread ID, which is now going to pass into Archon for all future calls. And so it's going to be using this API endpoint that I created for Archon, um, since that first video as well. And so it's making that first request already. You can see that Wind Surf is waiting for the API request to finish.

And this will take a little bit because Archon has that full workflow where it goes through the Reasoner model, create that initial scope for the agent, and then it goes to the Coder agent, like we see right here, that performs RAG to get the Pantic AI documentation, combines it all together to create that initial version of our AI agent.

And there we go. It used the code from Archon to build us agent.

An AI Agent that BUILDS AI Agents

so those are our two tools let's go
ahead and send in this request so the
first thing it'll do is call this tool
to create our thread ID which is now
going to pass into archon for all future
calls and so it's going to be using this
API endpoint that I created for archon
um since that first video as well and so
it's making that first request already
you can see that wind surf is waiting
for the API request to finish and this
will take a little bit cuz archon has
that full workflow where it goes through
the Reasoner model create that initial
scope for the agent and then it goes to
the coder agent like we see right here
that performs rag to get the pantic AI
documentation combines it all together
to create that initial version of our AI
agent and there we go it used the code
from archon to build us agent.

Get Our Mobile App

An AI Agent that BUILDS AI Agents

Transcription