README.md

Agentica

The simplest Agentic AI library, specialized in LLM Function Calling.

Don't compose complicate agent graph or workflow, but just deliver Swagger/OpenAPI documents or TypeScript class types linearly to the agentica. Then agentica will do everything with the function calling.

Look at the below demonstration, and feel how agentica is easy and powerful.

import { Agentica } from "@agentica/core";
import typia from "typia";

const agent = new Agentica({
  controllers: [
    await fetch(
      "https://shopping-be.wrtn.ai/editor/swagger.json",
    ).then(r => r.json()),
    typia.llm.application<ShoppingCounselor>(),
    typia.llm.application<ShoppingPolicy>(),
    typia.llm.application<ShoppingSearchRag>(),
  ],
});
await agent.conversate("I wanna buy MacBook Pro");

shopping-chat.mp4

Demonstration video of Shopping AI Chatbot

Guide Documents

🏠 Home

Preparing detailed guide documents.

Until that, please satisfy with README document of each module.

📖 Features

• Core Library
  • Introduction
  • How to Use
    • Setup
    • Chat with Backend Server
    • Chat with TypeScript Class
    • Multi Agent Orchestration
  • Principles
    • Agent Strategy
    • Validation Feedback
    • OpenAPI Specification
• Benchmark Program
  • Introduction
  • How to Use
    • Setup
    • Server Application
    • Client Application
• WebSocket RPC
  • Introduction
  • How to Use
    • Setup
    • Server Application
    • Client Application
  • NestJS Application
    • Bootstrap
    • API Controller
    • Software Development Kit
  • Principles
    • Remote Procedure Call

Appendix

• API Documents
• ⇲ Benchmark Result

Roadmap

Guide Documents

In here README document, @agentica/core is introducing its key concepts, principles, and demonstrating some examples.

However, this contents are not fully enough for new comers of AI Chatbot development. We need much more guide documents and example projects are required for education. We have to guide backend developers to write proper definitions optimized for LLM function calling. We should introduce the best way of multi-agent orchestration implementation.

We'll write such fully detailed guide documents until 2025-03-31, and we will continuously release documents that are in the middle of being completed.

Playground

https://nestia.io/chat/playground

I had developed Swagger AI chatbot playground website for a long time ago.

However, the another part obtaining function schemas from TypeScript class type, it is not prepared yet. I'll make the TypeScript class type based playground website by embedding TypeScript compiler (tsc).

The new playground website would be published until 2025-03-15.

Optimization

As I've concenstrated on POC (Proof of Concept) development on the early stage level, internal agents composing @agentica/core are not cost optimized yet. Especially, selector agent is consuming LLM tokens too much repeatedly. We'll optimize the selector agent by RAG (Retrieval Augmented Generation) skills.

Also, we will support dozens of useful add-on agents which can connect with @agentica/core by TypeScript class function calling. One of them is @wrtnlabs/hive which optimizes selector agent so that reducing LLM costs dramatically. The others would be OpenAI Vector Store handler and Postgres based RAG engine.

With these @agentica/core providing add-on agents, you can learn how to implement the Multi-agent orchestration through TypeScript class function calling, and understand how @agentica/core makes the Multi agent system interaction super easily.