AgentKit

Powerful Toolkit To
Build, Deploy, And Manage
AI-Agents for your App.

AgentKit is a powerful tool suite that simplifies the creation and deployment of production ready Agentic AI systems or AI Agents. It provides a range of features such as built in memory and document retrieval via vector dbs, powerful tools, webhooks, cron jobs etc.

Brought to you by Simon Weniger

Agents

Create Custom LLM Agent

An agent is a program or system that is designed to perform a specific task or set of tasks. In the context of Supergent, an agent is a Large Language Model that has access to a suite of tools and can decide which tool to use based on user input. There are two main types of agents: action agents and plan-and-execute agents. Action agents decide on an action to take and take that action one step at a time, while plan-and-execute agents first decide on a plan of actions to take and then execute those actions one at a time. An agent can be invoked with memory as well as a document which gives the Agent the possibility to remember previous messages and do Question/Answering.

What you can Build

  • QA over your own data, documents or backends
  • Custom GPT-4 powered support chatbots that now your business and customers
  • AI empowered product customization.
  • Productivity tools that are always right. No more haluzination or outdatet info.
  • AI assistants that can understand and carry out complex tasks.

Documents

Add Documents to Agents

AgentKit allows users to do question answering over documents by attaching a Document to an Agent. Documents are files, such as PDF, TXT, images, Markdown etc. that can be ingested by passing a publicly available URL to the AgentKit API. AgentKit splits these documents into smaller chunks and stores them in a vector database for use downstream. This approach has some downsides when working with tabular data. We are looking into how to best approach this problem. This is a community-made SwiftSDK for the Superagen LLM Framework. You need to have a superagent API Key in order to use the SDK.

What Documents can an Agent use?

  • The document type could be either a file like a Text, Pdf, or Csv.
  • Or it could be a Publicly available link like a YouTube Video, Url.
  • The Documents can be created and added to an agent by yourself
  • Your users can extend the Agents knowledge by adding their own documents

Prompts

Custom Prompt Tamplates

AgentKit allows users to configure their Agents using prompts. A prompt is piece of text that gives context to the LLM. It can contain instructions on how the Agent should act, input variables used for injecting data into the prompt. Common examples of input_variables are human_input, question and chat_history for in context memory.

What you can do with Prompts?

  • Pre configure the LLM with a optimized prompt for your usecase.
  • With prompt tamplates you can give the LLM context of previous conversations
  • Give the LLM context of your user as well as its task.
  • Allows powerful answer throught simple questions.

Tools

Extend the Agent Capabililties

This guide will get you all set up and ready to use AgentKit Tools. AgentKit allows users to configure their Agents using tools. A tool is a Api or piece of code that helps the LLM to validate its response. It can be an api that AgentKit calls to get real time information or a Zapier integration. Tools are helpfull to give LLMs access to real time and real world data while also reducing haluzination

Available Tools are?

  • Wolfram exper level knowladge database.
  • Websearch to connect your agent to the web.
  • Replicate so you can create multi modal agents.
  • Zapier allows you to build personal assistants and AI powerd automation.
  • Other Agent. You can connect Agents with each other.
  • OpenAI Plugins so you have access to the biggest AI ecosystem.