AI Data Scientist for
  • Jupyter

    Jupyter
  • Metabase

    Metabase

A Chrome Extension that operates your analytics apps for you. MinusX is the fastest way to get insights from data.

Explore Data in Natural Language

Generate hypotheses and manipulate data

Jupyter

Metabase

Help is ⌘+k away

Interop with MinusX to modify or extend existing notebooks

Jupyter

Metabase

Select & Ask

Select an area and ask questions, or ask for modifications

Jupyter

Metabase

Built For

live

coming soon

on roadmap

jupyter

jupyter

metabase

metabase

posthog

posthog

tableau

tableau

grafana

grafana

google colab

google colab

google sheets

google sheets

google slides

google slides

amplitude

amplitude

Want us to hurry up, or don't see your favourite tool? Let us know!

try minusx today!

FAQs

Have a different question and can’t find the answer you’re looking for?

Join minusx Discord

Is minusx a new platform?

No! MinusX works in your existing analytics tools like Jupyter notebooks, Metabase, etc. Given an instruction, MinusX operates the app - by clicking and typing, just like you would - to analyze your data and answer queries.

How is my data handled?

We store all metadata on your browser. This includes table schemas, code, queries, and a sample of query outputs. This is used to set the context for your instruction and sent to the LLM. You can find a non legal-speak overview of the data we collect in our privacy policy

What can this do, over LLM chat apps?

MinusX seamlessly integrates with existing tools. This means that you never have to get out of your workflow to answer questions. MinusX also has complete context of the app you're using, and can choose the right action for the right context.

How much does this cost?

The product is currently free to use. In the future, we'll charge a monthly subscription fee, and support local models/bring-your-own-keys.

Which models do you support?

Currently we support Claude Sonnet 3.5, GPT-4o and GPT-4o mini. We are also working on a way to let you choose your own models as well as bring your own models.

What's the plan here?

We want to build the best AI data scientist - one that uses whatever tool and context is needed to get the job done. We think the path to truly autonomous agents runs through useful, specialized intermediaries.