Build Intelligent Autonomous Agents
Assemble and train skill-based agents from composabl building blocks.
pip install composabl
from composabl import Agent, Skill, Sensor, Scenario, Runtime
from .teaching import IncrementTeacher
from .scenarios import increment_scenarios
from .sensors import sensors
from .perceptors import perceptors
increment_skill = Skill("increment", IncrementTeacher, trainable=True)
for scenario_dict in increment_scenarios:
scenario = Scenario(scenario_dict)
runtime = Runtime(config)
agent = Agent(runtime, config)
Introducing the Composabl SDK.
Democratizing the power of Autonomous Agents for everyone!
Focus on teaching skills, not writing code.
Define the scenarios where your agent will need to succeed and the skills it will need to accomplish the task, all in just a few lines of code.
Use familiar tools to build your agents.
Quickly install our Python SDK and start creating your autonomous agent through a familiar language.
Train agents locally, on-premises, or scale on the cloud.
You shouldn't have to build a cloud app or manage software containers to train intelligent agents. Now you don't have to.
Leverage SME knowledge
Utilize the knowledge of your experts, and integrate it into autonomous agents.
Bring your own simulator
There is no need to acquire a new simulator. Composabl is built to integrate and scale with your existing data and simulation tools
Make your agent explainable
Track the behaviour of your agents and compare them against each other to identify the best performing agents.
Frequently asked questions
Here are some answers to the questions we most frequently hear about Composabl, reach out to us.
What is an intelligent autonomous agent?
An intelligent autonomous agent controls systems and processes by perceiving and acting (making decisions) in real-time.
What are some things that I can do with intelligent autonomous agents?
Intelligent autonomous agents can control equipment like CNC machines, bulldozers, drones, extruders, robotic arms, and chillers or processess like scheduling, trading, and maintenance.
What is the difference between an intelligent autonomous agent and a control system?
Intelligent autonomous agents have advantages over traditional control and optimization systems. They provide more nuanced, human-like decision-making. We wrote a blog post about it if you'd like to learn more.
What are the building blocks for intelligent autonomous agents?
Skills! You can use Composabl to define a skill by programming the logic for that skill, asking Composabl to learn the skill, or importing a skill as an existing control or optimization algorithm.
Do Composabl intelligent autonomous agents replace my automation technology?
No. Composabl allows you to combine existing controls and optimization technoglogy with expert rules, machine learning, and AI to augment and outperform existing systems.
Do Composabl intelligent autonomous agents replace my human experts?
Not really. Most people use Composabl agents to advise, train, and augment their experts. Some applications require full autonomy and Composabl supports that.
Can't I do all this with large language models (LLM), like AutoGPT?
Large language models cannot be trusted for most real-world decisions. Use them anywhere you'd follow instructions that you find on the internet verbatim. We wrote a blog post about this too.
What is Machine Teaching?
Machine Teaching is our methodology for building intelligent agents from skill-based components. We believe that the paradigm in autonomy is shifting from programming to teaching. We wrote a book about it if you'd like to learn more.
Who is on the team at Composabl and how do you know how to build a platform for intelligent autonomous agents?
We're a group of eningeers, programmers, and data scientists that have been designing, building, and deploying intelligent autonomous agents for some of the largest companies over the last 5 years.
Why are you charging for your product; what's so hard about building intelligent autonomous agents?
You'd be surprised how difficult it is to train an intelligent autonomous agent from scratch and how much code you'd have to write if you did. The value in the Composabl platform is all of the things that we take care of under the hood so that you can build valuable agents quickly and easily.
Can I try the product for free?
Yes. We offer a free 30-day trial for you to build as many agents as you'd like. During the free trial, you can train agents on your local machine or on-premesis servers.
What benefits do I get by upgrading to the small teams version of the Composabl?
The biggest advantage is the ability to train your intelligent agents at massive scale on the Microsoft Azure Cloud.