I made MCP (Model Context Protocol) alternative solution, for OpenAI and all other LLMs, that is cheaper than Anthropic Claude

Community
, , ,
1

Detailed Article

I made MCP (Model Context Protocol) alternative solution, for OpenAI and all other LLMs, that is cheaper than Anthropic Claude - DEV Community

I made MCP (Model Context Protocol) alternative.

My alternative solution is utilizing LLM Function Calling feature, provided by Swagger/OpenAPI and TypeScript class functions, enhanced by compiler and validation feedback strategy. With these strategies, you can fully replace the MCP of Anthropic Claude to much smaller models like gpt-4o-mini (Possible to replace to Local LLMs).

Below is a demonstration of my alternative solution, searching and purchasing product in a shopping mall. Its bacakend server is composed with 289 number of API functions, and I could call all of these API functions just by conversation texts in the gpt-4o-mini model of 8b parameters.

import { Agentica } from "@agentica/core";
import { HttpLlm, OpenApi } from "@samchon/openapi";
import typia from "typia";

const agent = new Agentica({
  service: {
    api: new OpenAI({ apiKey: "*****" }),
    model: "gpt-4o-mini",
  },
  controllers: [
    HttpLlm.application({
      model: "chatgpt",
      document: OpenApi.convert(
        await fetch(
          "https://shopping-be.wrtn.ai/editor/swagger.json",
        ).then(r => r.json())
      )
    }),
    typia.llm.application<ShoppingCounselor, "chatgpt">(),
    typia.llm.application<ShoppingPolicy, "chatgpt">(),
    typia.llm.application<ShoppingSearchRag, "chatgpt">(),
  ],
});
await agent.conversate("I wanna buy MacBook Pro");
3 Likes