Launching today
Databench by Alkera
Open-source collaborative agentic workspace for data teams
342 followers
Open-source collaborative agentic workspace for data teams
342 followers
Databench by Alkera is the open-source, multiplayer workspace for data science, analytics, and engineering. Collaborate live alongside teammates and agents in notebooks and chats, run any cell or agent on your laptop, another computer, or a GPU node and launch many agents in parallel to explore ideas. Every result traces back to the data and code behind it. Try it at https://alkera.ai with a generous free tier or host Databench yourself (https://github.com/AlkeraAI/Databench).
Hey Product Hunt! 👋 I'm Rick, co-founder of Alkera, building Databench with Andrew Tran and Tony Li. Before starting Alkera, we worked closely with data teams at places including TikTok, Ramp, and Hudson River Trading.
Alkera provides a unified platform for data engineering, data science, and data analytics work. We've noticed a significant need for an agent-native data science solution beyond traditional Jupyter notebooks.
❌ The Problem
While new agent-native interfaces such as ADEs have significantly increased software engineers' productivity, data teams are still stuck with the traditional Jupyter notebook, with little evolution or adaptation for agent-first, collaborative workflows. Another problem with agentic work for data is the lack of accountability, traceability, and reproducibility, which makes it hard to verify and reproduce agent-generated work.
This is why we built Databench, the open-source multiplayer workspace for data science, analytics, and engineering. Work alongside teammates and agents in notebooks and chats and run any cell or agent on your laptop or a remote computing node.
✅ What Databench provides:
Open-source subset of the Alkera product licensed under Apache 2.0
Multiplayer SQL and Python notebooks that you share live with teammates and agents
Connection to databases that are SQLAlchemy compatible (PostgreSQL, SQLite, MySQL, and many more)
Reactive notebooks based on marimo that rerun stale cells based on changes to SQL & Python
Fully web-based with rich file editors and notebook editing
Live collaboration with live cursors and edits from teammates and agents
Run chats and notebooks anywhere: your laptop, a remote compute node, or a GPU cluster
Declarative charting library for pretty graphs every time in your notebooks
Try it free at https://alkera.ai or self-host from GitHub: https://github.com/AlkeraAI/Databench
If you have any questions or concerns, the team will be online all day to help!
@rick_alkera Love that stale cells rerun by themselves. I always forget to rerun something in Jupyter and then wonder why my chart looks off..
Hi, I'm the CTO of Alkera and led the development of Databench.
First, real-time collaboration on notebooks has always felt fragile, while research should be collaborative. I've lost count of the times I've tried collaborating on a Jupyter notebook over VS Code Live Share, only to hit desynced cells. That experience gets worse when an agent joins in and starts making changes too. With Databench, we made multiplayer a core part of the product. Second, I rarely run any interesting data analysis on my laptop, especially for anything that involves a GPU. So, I wanted an experience that was designed from the ground up for interacting with remote compute.
I'd love to hear from anyone who has fought the same battles messing around with data, especially in notebooks.
Hey everyone! I'm Tony, one of Alkera's cofounders along with Rick and Andrew.
Before Alkera, I worked on a trading project where I pulled odds from sportsbooks and prediction markets to find +EV bets and arbitrage. My data pipeline lived in Python files and warehouses, and although it ran, I couldn't see how data moved, nor did I trust my coding agents, which were great at SWE, but not no much at data that lived across different tools.
Databench is the workspace I wish I had and use now. Would love to see how others use it!
@rick_alkera working with teammates and AI agents in the same workspace sounds really useful. Exited to see how people use databench.
@rohit62661 Thanks for the support! Would love to hear your feedback!
Congrats on the launch! I like that the data and code stay visible alongside the agents' work. Much easier to discuss an answer when everyone can see how it was reached.
Congrats on the launch, Rick!
@michelle_marcelline Thanks for the support! Love the work you're doing at Promptingco
This is super cool!