Developers
Build with confidence.
Use Howso Platform to:
- Analyze your data
- Capture exact relationships behind every output
- Create predictions with a simple csv, a database, or wrap entire ML models
All with full attribution. This means being able to point to the exact data used to generate predictions or pieces of content. It’s open source and great for people and teams who care about how data is used.
Open source AI/ML technology perfect for data centric applications.
Pre-Reqs
Python
Proficiency in Python is necessary to get the most out of our tools. You’ll use Jupiter Notebooks to load, train, and analyze data.
Basic Understanding of Data Science or Machine Learning
We recommend a basic understanding of data science or machine learning. Our documentation aims to reduce this barrier, but having this understanding will reduce the learning curve and increase productivity.
Bring Your Own Data
Howso requires you to provide your own dataset to train and analyze. We work best with tabular structured data.
Requirements
Hardware
Howso supports machines running modern versions of Windows, Linux, & Mac operating systems with at least 8GB of RAM. Our algorithms are CPU-based and the more cores you have, the more performant your system will be.
Note: GPU accelerators are not required or suggested as they do not provide performance gains with Howso’s technology.
Software
Option 1: Using a Pre-built Dev Container
- VS Code
- Docker
This is the fastest and easiest way to get up and running without the need to manage individual packages.
Option 2: Local Installation
- Howso Engine + Dependencies
- Local Python Environment
- Application for running notebooks (e.g. IDEs, Jupiter server, etc.)
If you want to do a local installation with PiPy, this is the route to for you.
License & Usage
Howso’s software is offered via an OSI and FSF approved open source license. This is a copyleft license which grant permissions to use while restricting proprietization and distribution.
Playground
Examples?
Yeah, we’ve got them.
We’ve created a number of scenarios on our playground that show Howso in action. These also explain how our ML/AI works.
Open Source
We’re On a Mission
In September 2023, we open sourced the Howso Engine to share our technology and provide transparency into how it works. This makes our software readily available to developers, engineers, data scientists, and anyone interested in building solutions with Howso. We hope to grow a global community of developers who contribute to improving and expanding Explainable AI.
Ready to get cracking?
Integrations
Use Your Existing Data
Howso has a number of integrations to help you get up and running in your existing AI Platform of choice.
Databricks
Snowflake
Azure Studio
Resources
Documentation
Guides
Our documentation contains a couple of guides: Getting Started & User Guides. The Getting Started guide cover the basics of installation, key concepts, and terminology. The User Guides focus on explaining capabilities and core concepts vital to working with Howso’s technologies.
Recipes & Datasets
Our recipes contain pre-written Python snippets to help you start using Howso. Recipes are organized to help illustrate common workflows and save you time. We’ve also curated a group of open source datasets that are used in various recipes or can be used to explore other capabilities.
API Reference
Unlock the full potential of your applications using our API reference. Here you’ll find details on the various methods and parameters that can be used when writing your own applications. Our API empowers developers to connect, enhance, and scale their projects effortlessly.