FireKit AI
  • Welcome to FireKit AI
  • Guide
    • Introduction
    • Core Concepts
      • AI Chains
      • Blocks
      • Edges
    • Rules
    • Debugging
  • Blocks
    • Agent
    • Skill
    • VectorStore Reader
    • Text
    • Chat Memory
Powered by GitBook
On this page
  • Input/Output
  • Agent
  • Skill
  • LLM
  • VectorStore Reader
  • Text
  • Chat Memory
  1. Guide
  2. Core Concepts

Blocks

An introduction to FireKit blocks

PreviousAI ChainsNextEdges

Last updated 1 year ago

FireKit blocks are functional components within an application. You can chain blocks together to implement complex logic and powerful AI capabilities.

Input/Output

These blocks represent the input and output of the application. The chain will begin with the input and end with the output.

Agent

An agent is an AI entity capable of decision making. Given a goal, it undertakes a series of actions, observing and reasoning as it goes, until it achieves its intended outcome.

The agent operates on a Large Language Model (LLM), which can be invoked multiple times during a single execution of the agent block. By pairing specialized LLM prompts with conventional programming techniques, the system is guided to function as an agent.

Read the section for more details.

Skill

The skill and the agent are tightly integrated. The skill provides functionality to the agent, while masking the intricacies of its underlying implementation. Multiple skills can be connected to an agent, from which the agent can choose to achieve its objective.

You implement the functionality of a skill with a chain contained within the skill block. The blocks within a skill chain are only executed when an agent decides to employ the skill.

Read the section for more details.

LLM

This block has the functionality of a traditional Large Language Model. Given a text input, it will generate a text output.

VectorStore Reader

This block provides an interface for querying from a vector store and fetching the text representations of matching vectors. This is useful for fetching information relevant to a provided text.

Text

The Text block serves as an interface for formatting text within a chain. It can transform text outputs from preceding blocks and also supply text as input to subsequent components in the chain.

Chat Memory

The Chat Memory block provides the history of inputs and outputs from chain executions as text. This serves as a mechanism for persisting memory of the conversation among executions of the chain.

Read the section for more details.

Read the section for more details.

Read the section for more details.

Agent
Agent
VectorStore Reader
Text
Chat Memory