AI meets spreadsheets
AI for 1 billion+
1 in 7 humans use Excel or Google Sheets.
ExploreUp to 100+ times faster sampling. Random sample 1 billion+ rows down to 1 million in only a few seconds. Switch between data engines if you need (Pandas, DASK, Spark, GPUs with RAPIDS). Use S3 like a local drive on your laptop with simple copy/paste operations. Query Databases using SQL.
Easily create charts and profile your data with one click. Quickly understand how your data is distributed and identify hard-to-find correlations.
Process, clean, and prepare your data without any coding. When clicking is not enough, use 200+ Excel compatible formulas. Add custom Python code using the built-in editor for more complex requirements. Export the final clean data to a file, data lake, or database directly from the app. To build repeatable data pipelines, export to pure Python code, and run anywhere you need.
You can also use Excel for data preparation! (only at Practicus AI). Share pure Excel or Google Sheets with others, so they can analyze the data and make changes, and send it back to you. Practicus AI can then capture those changes so you can apply them to any data locally or in the cloud, or export to Python code.
Build AI models on past data with one-click using Automated Machine Learning (AutoML) and then make predictions on unseen data. If you are a data scientist, or working with one, export to Jupyter code and share your experimentation details in a central database (MLflow). Your team will save a significant amount of time by avoiding coding from scratch and by collaborating better.
Predict the unknown with one-click. Search and use models built with any technology / platform. You can use model APIs hosted anywhere, or use models cold stored on your laptop, S3 or other location, which will become live within seconds. With a short startup time and ease of use, your organization’s ML models will reach new users easier.
Excel PredictYou can also use pure-Excel for prediction! (only at Practicus AI). Export and embed your AI models into Excel or Google Sheets, attach to an email, and allow your users to make predictions completely offline with no app / plugin / macro required. You can use the AutoML models, or export your own custom models. Currently supporting Linear Regression, Logistic Regression, SVM and Decision Trees.
OperationsCreate production-ready data processing code with one-click, embed into Airflow or other orchestration engines. Choose your preferred data processing engine including Pandas, DASK, RAPIDS (GPU), RAPIDS + DASK (multi-GPU) and Spark. Create cloud IAM policies that apply to all of our users to create audit trails for data access and processing. Easily roll out MFA or SSO for all of your users.
Cut your data preparation time +50%, code only when needed
Export simpler models to pure Excel to increase user base
Simplified model search and one click visual inference improves consumption of your models
Explore various data sources to analyze raw or structured data
Build AI models and predict the future, export to Excel to share
Easily join technical experts for complex AI projects, where your expertise is needed the most
Easier and faster ETL with spreadsheet interface
Easily switch between Pandas, DASK, RAPIDS, Spark and more
Export data processing steps into pure Python and embed into Airflow or other orchestrator with one-click
Activate Cloud (optional)
Some of our advanced features require cloud workers.
You can activate in 2-5 minutes and use the forever free cloud tier!
Takes ~3 minutes. This is a one-time task for all users.
If you do not have an Amazon Web Services (AWS) cloud account, please click here to create one for free.
After AWS account creation, you will also need a separate admin user for security reasons.
Please check our step-by-step guide for help on AWS account creation.
Takes ~2 minutes. This is a one-time task for all users.
Please click here to enable (subscribe to) our offering on AWS marketplace. This step will also unlock the forever free cloud tier for 2 vCPUs and 1 GB RAM (t3.micro). When you prefer to use larger cloud capacity, AWS simply bills you for the hours that the cloud node is running.
If you plan on using advanced Nvidia GPUs, please click here to enable GPUs as well.
Once your subscription is active, you do not need to continue on the AWS marketplace to configure. You can safely close the page since the below step will take care of configuration.
Takes ~10 seconds.
Please open the Practicus AI app, go to settings page and enter your AWS access key information. Confirm the verification is completed on the same page, and you are ready to go!
You can use pip as an alternative installation method.
You can pip install practicus (Windows/macOS: Python 3.7 – 3.10, Linux: 3.8 – 3.10) and then run practicus from the terminal. Or run, python -c “import practicus; practicus.run()” (python3 for macOS or Linux).
Installing using pip will give you the exact same end-to-end GUI application experience. Similarly, if you download the packaged app you can still code freely when you want to. So, please select any method of installation as you prefer.
As for any Python application, we strongly recommend you to use a virtual environment such as venv or conda. Please check the recommended QuickStart scripts on this page to create a virtual env, install Practicus AI and run with peace of mind and in one go.
For server environments or API only usage, you can pip install practicuscore to install the core library by itself without the GUI elements. (Linux: Python 3.6 – 3.10, Windows/macOS: Python 3.7 – 3.10)
This is a small library with fewer requirements and no version enforcement for any of its dependencies. It’s designed to run in existing virtual environments without overriding the version requirements of other libraries. Please check the documentation for more details.
:: install python -m venv %UserProfile%\practicus\venv %UserProfile%\practicus\venv\Scripts\python -m pip install --upgrade practicus :: run %UserProfile%\practicus\venv\Scripts\activate practicus
# install python3 -m venv ~/practicus/venv ~/practicus/venv/bin/python -m pip install --upgrade practicus
# run source ~/practicus/venv/bin/activate practicus
Instead of running Practicus AI from the command prompt (terminal) you can create shortcuts and run with double click. The below tips assume you installed using the QuickStart tips above.
Windows: Navigate to %UserProfile%\practicus\venv\Scripts\ folder and locate practicus.exe, which is essentially a starter for practicus Python library. You can right click and select pin to start. You can also create a shortcut to this .exe and change its name to Practicus AI and its icon by downloading our icon practicus.ico.
macOS: You can download Practicus AI Starter app which is a tiny (100KB) app that starts the Python virtual env in ~/practicus/venv , and then starts Practicus AI GUI from practicus Python library. To keep the app in dock please drag & drop the .app file on the dock itself. Right clicking and choosing “keep in dock” will not create a shortcut.