Imagination Engine is an agentic orchestration framework for rapidly building agentic architectures in Python. It accelerates agentic development, empowering developers to craft architectures that enable LLMs to excel in higher-order tasks like idea generation and critical thinking.
This library is designed to be used as a backend for AI apps and automations, providing support for all major LLM providers and locally-hosted model endpoints.
This project is currently in development. More features and support coming very soon.
This will ultimately be a one-liner once the library is published. For now import the following into your python script:
from agent import Agent
from graph import Graph
from utils.start_end import START, END
Done in the format standard to your API.
Initalize the agents you'd like to call with a simple schema:
agent = Agent(
client=client,
model="gpt-4o",
name="Agent",
system_prompt="You are a friendly assistant.",
description="This agent is an example agent."
)
Graphs connect agents together using nodes and edges, routing the execution flow and all needed information through the graph. Each node represents an agent and each edge represents a conditional under which that agent is called. This conditional can be defined in natural language for each agent, within its system_prompt.
In the case of multiple edges branching from one node, agents can understand their routing options using the description field of connecting nodes.
Every graph accepts user input at the START and returns a response to the user at the END.
With this basic understanding, a graph can be created in just a few lines.
graph = Graph()
graph.add_node(agent)
graph.add_node(summarizer_agent)
graph.add_edge(START, agent)
graph.add_edge(agent, summarizer_agent)
graph.add_edge(summarizer_agent, END)
You're done! Prompt your agentic architecture.
graph.invoke("Hello there!")