Keptune

What it does

Keptune AI helps scientists and researchers analyze their data by automating tedious data-cleaning tasks, running statistical tests, making charts, etc. using natural language prompting. It offers a simple yet powerful alternative to apps like Excel, SPSS, Tableau, Prism, and Origin. The idea for this app came out of my professional experience as a Scientific Software Engineer who wrote custom apps and scripts for scientists to save them several hours per week by automating repetitive data cleaning and analysis tasks. With the help of Gemini API, Keptune aims to make it possible for every scientist to write custom data analysis workflows tailored specifically to their needs.

Keptune AI talks to the Gemini API on the backend to plan the analysis steps, write Python code to accomplish each step, and offer suggestions for follow-up analysis. Because the code is generated on demand, thanks to the Gemini API, there is infinite flexibility in the types of analyses that can be automated. Users also have the option of editing and re-running any code generated by the Gemini model. This combination of natural language prompting along with the code editing capability makes it an ideal tool for both beginners and experts. The python code is run in the browser using Pyodide (a port of CPython to WebAssembly) so there is no need to provision virtual machines for users, making it cost-effective to run this app at scale.