The problem

After I sold my last company, I knew I wanted to work in web3. I’d followed bitcoin from its infancy and hearing Ethereum described as “programmable money” clicked with me immediately. The eureka moment was when I realized you could deploy a smart contract that acts as a credit union. And you could do it in a weekend. I dug in deep, and the natural place for a data nerd to start was getting a ton of rows to play with in a Jupyter Notebook.

But surprisingly–no one wanted to help me do that. I couldn’t find a single service where I could use the lingua franca of data (SQL) and get my hands on billions of rows of data. I started toying with some other projects, like trying to predict the likelihood a wallet would transact with Tornado Cash in the next 30 days. But alas, I couldn’t get the data to do it. I found crypto data in BigQuery, but I learned I’d need to first download all the data and decode it. And it was going to cost thousands of dollars to run those queries. And if I missed a column, I’d need to pay again. And I needed to find a place to store it. And I needed to find a place to run the pipelines…

I found a problem: it’s really hard to get reliable web3 data at scale. You’re either using community built views an anonymous user created in Dune and hoping you can download it all in a CSV, or building data pipelines from scratch. I know this because I heard it over and over from 100+ founders, engineers and data scientists in web3. The DIY route is crazy expensive:

• Hire 2 data engineers ($500k+ per year)
• Run a node on Quicknode or Alchemy ($20k+ per year)
• Pay a cloud provider to run your custom pipelines ($20k+ per year)
• Pay BigQuery or Snowflake for compute and storage ($20k+ per year)
• Spend countless hours modeling the data
• Watch everything break after “the merge” and stay up till 4 AM rewriting your ETL

The services for web3 data aren’t much better. Using the “unofficial” Dune Python API? Good luck getting 10,000 rows out, let alone 1 million.

It’s a mess, but I like cleaning up. And I just spent the last 10 years neck-deep in the modern data stack. First, as a practitioner using it, then as a founder building it.

I cracked open VS code and started hacking on a little side project: Luabase. I originally called it “moonbase,” but moonbase.com was a bit out of reach. “Lua” is moon in Portuguese and the .com was $12. Sold.

I built the first version in 30 days and started getting people to use it. Quickly, folks from a range of companies validated my approach, from developers at thirdweb to data scientists at leading crypto and VC firms.

The solution

Luabase is the modern data stack for web3. We built the data stack of your dreams. The one you’d build if you had the time and engineering capacity to build it yourself. We make it easy to query blockchains with SQL, analyze it in notebooks, and embed the data in your application.

Clean, fast and decoded data

Luabase ingresses data from blockchains and stores it in a highly optimized schema so you can get answers with sub-second performance, even across billions of rows. We also decode every log we can get an ABI for, so you can get the full story behind any on-chain transaction.

Notebooks to analyze the data

We have a form factor to interact with the data you’re all used to – Notebooks. You can iteratively build datasets and quickly pivot and visualize your data. Notebooks include Markdown cells so you can document your code as you build and collaborate with your team.

API and Destinations

You can access any of your datasets via API to embed data in your application or send the data to any of the destinations we support (S3, Snowflake, Google Sheets, Discord, Slack, etc.).

Where we're headed

The initial 30-day speed run of Luabase was just the beginning. While we tackled Ethereum first, we quickly added Polygon, Avalanche, Fantom and Bitcoin soon after. In the next few months we’ll be adding every EVM chain and Solana. We're already API-first but our design partners (like thirdweb, OpenZepplin and Artemis) told us they needed tools to make Luabase fit into their development workflow, so we'll be launching a Github integration next month and an SDK shortly after.

The seed round

This is a big problem; it will take money and incredible people to solve. In the earliest conversations with founders in web3, they not only wanted to use Luabase but to invest in it. So we raised $4.5m from a crew of web3 founders and investors to build Luabase with us:

• Costanoa Ventures
• 6MV (6th Man Ventures)
• Furqan Rydhan (Founder @ thirdweb)
• Shaan Puri (Founder @ Milkroad)
• Jon Ma (Co-founder @ Artemis)
• Jack Herrick (Founder of WikiHow)
• Josh Hannah (Former GP at Matrix)

Join us

We’ve assembled a team of world-class software and data engineers. If you’re a barber for hairy data problems, drop me a note at mike@luabase.com and join us in building the future of data in web3.