Using Plugins Without ChatGPT

GPT builders Plugins / Actions builders
1

I am looking at trying to use the retrieval plugin as model memory in a gpt4-based custom application. Is there a current flow for doing something like that, or has anyone tried something similar?

Thanks!

1 Like
2

Plugins aren’t available through API, just Chat GPT. you can look at projects like LangChain that accomplish similar things through API.

2 Likes
4

Is there any plan to dd this in api call as parameter e.g. model name gpt-3.5-turbo, useplugin=true, plugins enabled = plugin1, plugin2 etc.

Who can share roadmap ?

5

This is all we have:

6

You can use lang chain and create an agent and add your plugin here is an example:
import { ChatOpenAI } from “langchain/chat_models/openai”;
import { initializeAgentExecutorWithOptions } from “langchain/agents”;
import {
RequestsGetTool,
RequestsPostTool,
AIPluginTool,
} from “langchain/tools”;

export const run = async () => {
const tools = [
new RequestsGetTool(),
new RequestsPostTool(),
await AIPluginTool.fromPluginUrl(
“-- plugin link”
),
];
const agent = await initializeAgentExecutorWithOptions(
tools,
new ChatOpenAI({ temperature: 0 }),
{ agentType: “chat-zero-shot-react-description”, verbose: true }
);

const result = await agent.call({
input: “what t shirts are available in klarna?”,
});

console.log({ result });
};