MinusX is a chrome extension that adds a sidechat to your analytics apps like Jupyter and Metabase. Given an instruction, our agent operates the tool - by clicking and typing - to analyze data and answer queries.
Hey Product Hunt! 👋
We're @nuwandavek , @ppsreejith and @arpit_tarang , and we're building MinusX, an AI Data Scientist for Jupyter and Metabase. MinusX is a Chrome extension (https://minusx.ai/chrome-extension) that adds an AI sidechat to your analytics apps. Given an instruction, our agent operates your app (by clicking and typing, just like you would) to analyze data and answer queries. Broadly, you can do 3 types of things:
- 🔮 Ask for hypotheses and explore data
- 🛠️ Extend and modify existing notebooks/dashboards
- 🔍 Select a region and ask questions about it
🤕 Why are we building this?
Developers just want answers. Product managers just want answers. Analysts want 10 clones of themselves. This is a sentiment we hear over and over. So what’re the options do people have?
1. Continue business as usual with 1000s of back and forths - 🥲 painful
2. Move to a new fancy AI analytics tool (means convincing all colleagues to also move to the new tool, abandoning all existing workflows) - 🥲 extra painful
3. Go to chatGPT which has no real context of your data, your past, your workflow and admonish it till it gives something passable - 🥲 less painful, but sad
If this sounds like you, you’ll love MinusX! 🚀🚀🚀
An AI Data Scientist is a Scientist, not yet-another-new-fancy tool. [🧑🔬 >>> 🛠️]
MinusX just works, where you already are! It uses the full context of your data, current workflows, and is deeply integrated into both Jupyter and Metabase. It "retrofits" analytics software with intelligence, something LLMs have begun to unlock. We also get a kick out of the fact that we use the same APIs humans use (clicking and typing), so we don’t really need "permission" from any analytics app 👀
The product is currently free to use. We’re working on supporting local models and bring-your-own-keys soon! We'd be super stoked for you to try out MinusX! You can find the extension here: https://minusx.ai/chrome-extension. We've also created a playground with data, for both Jupyter and Metabase, so once the extension is installed you can take it for a spin: https://minusx.ai/playground
We'd love to hear what you think about the idea, and anything else you'd like to share! Suggestions on which tools to support next are most welcome 🙏
❤️
Awesome platform! Apart from Metabase, will you be supporting more BI tools along the way? Can't help thinking of Tableau, Power BI, Google Data Studio (not too relevant)
@charlestehio Yep! We're currently exploring the tools mentioned in our home page (Grafana, Colab, Google Sheets, etc) but do tell us which tool you'd like us to add next at https://minusx.ai/tool-request
Love how simple the tool is and is targeting at some of the most widely adopted data tools like Metabase! I used to go back and forth between Metabase and ChatGPT all the time as I'm not super technical but need help with charts and data pull.
Congrats on the launch @nuwandavek and team!
This seems like a good tool for simplifying data analysis within existing workflows. Happy to see AI seamlessly integrated into Jupyter and Metabase, making data exploration and hypothesis testing so much more efficient. Congrats on the launch :)
This is a game changer for anyone juggling Jupyter and Metabase! 💡 I love that MinusX operates like a data-savvy sidekick, turning the tedious into the seamless. No more endless back-and-forths or migrating to complicated tools—just good, old-fashioned smart analytics with a twist! Can’t wait to see what features you guys roll out next. @nuwandavek, are you planning to integrate with any specific data visualization tools?
@nuwandavek@ema_elisi We're still exploring what tools to add support to (Currently looking at Grafana, Colab, Google Sheets etc). Please do let us know which tools you'd like us to add at https://minusx.ai/tool-request :)
@kavyaaatripathi Yes, it is somewhat similar, practically. But our approach allows for some really powerful experiences. Firstly, MinusX is the same interface that works on all your analytics apps. Your workflows across your analytics stack will remain the same.
Secondly, we’re working on cross tool support, which powerBI has no incentive of building. For example, you can import context for the project you’re analyzing from your google docs, or from an existing query on tableau or even from your email! You can analyze data in jupyter and import that into a presentation or a dashboard.
Wow, @nuwandavek, MinusX looks like an absolute game-changer! As a data analyst, I often find myself juggling between Jupyter and Metabase, and this AI assistant seems like a lifesaver. The ability to integrate directly into existing workflows without any hassle is just brilliant.
Can't wait to try it out and see how it boosts our data analysis efficiency! 🎉
This concept of an AI Data Scientist integrated into existing analytics tools like Jupyter and Metabase is quite intriguing. I'm curious about the potential limitations of MinusX in handling complex data queries compared to traditional methods. How does it ensure accuracy in the insights provided? Additionally, will there be plans for premium features or is the focus on keeping it entirely free? 🙏
@dance17219 We use a bunch of heuristics to handle complex queries which include the current query, sample outputs, and table schema as well as looking at existing dashboards and queries across your app which gives us useful context on picking the right tables for the query.
We're still working on the exact pricing strategy. We'll probably charge a monthly subscription fee and support local models/bring-your-own-keys.
MinusX sounds like a great addition for anyone working with Jupyter or Metabase. The idea of an AI Data Scientist that works within your current tools is super cool.
@oliviajanemitchell Yeah! We think the bottleneck really is intelligence and not the limitations of the tool themselves. So current tools should be able to go really far when augmented with MinusX!
Hey Vivek, Sreejith, and Arpit,
I'm wondering how you handle data privacy concerns, especially when working with sensitive information in Jupyter notebooks or Metabase?
Are there plans to expand support to other popular data analysis tools in the future?
Congrats on the launch!
@kyrylosilin Regarding sensitive information: 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. Furthermore, we're currently only using API providers (OAI + Claude) that do not themselves train on data accessed through APIs. Finally, we also plan to support local models / bring-your-own-keys in the future!
Regarding other data analysis tools, we're currently exploring the tools mentioned in our home page (Grafana, Colab, Google Sheets) but do tell us which tools you're most interested in adding and we'll prioritise accordingly: https://minusx.ai/tool-request
This is an awesome concept, @nuwandavek! The integration with Jupyter and Metabase is a game changer for analysts. 🔍 The ability to get answers without switching tools feels like a massive time saver. Excited to try out MinusX and see how it enhances my workflow!
@alexanderscottwilliams Yes! along with saving time, our current users also notice MinusX pickup on details they might miss in an analysis. Avoiding needless back and forths with their teammates is a definite positive too!
Hey @nuwandavek, MinusX sounds super useful, especially for someone like me who uses Jupyter all the time. One quick thing though—how well does it work with more complex dashboards? Just curious if there might be any hiccups with more intricate setups.
@nuwandavek@christopherdavidanderson We've tried it on some large notebooks and complex Metabase Questions. Do try it out and let me know if you've any issues, feedback, or requests!
@christopherdavidanderson I totally relate! In my previous job, I used to use Jupyter for like 80% of my data science use case! Do try it out and let us know if it falls short of expectations, we'll fix it! Also the current state of MinusX is only the beginning. We're working on some really exciting things for Jupyter which we'll share soon!
Congratulations on launching MinusX! 🎉
How does MinusX ensure that the AI Data Scientist accurately interprets the context of your data and workflows within Jupyter and Metabase?
MinusX