Announcing the Arraylake Community Tier

TL;DR

Today we’re launching the refreshed Arraylake Community Tier: a full-featured, free tier of the Earthmover platform. Anyone can now sign up and create their own repos, using either our free 10 GB cloud storage, or bring your own cloud storage at no extra cost. You can also serve your data through Earthmover Compute with 50 free compute credits per month.

There is no payment required, no sales call, and with 10 GB of cloud object storage included, no cloud account needed. If you’ve been waiting for a frictionless way to try cloud-native data workflows, this is it. Follow the quickstart and start building today!

From read-only to full-featured

When we launched the Arraylake Data Marketplace earlier this year, users with free accounts could browse the Marketplace and subscribe to public datasets. That is a great way to consume data, but we kept hearing a consistent theme from the community: You wanted to try out our data delivery APIs, test drive the catalog, and host your own datasets.

Now you can.

The refreshed Community Tier removes the wall between consuming data and producing it. Every Arraylake organization now gets access to the core platform, from ingesting data to serving up that data over different protocols. We believe the best way to understand what Arraylake can do for your team is to actually use it, end to end, with your own data.

What’s unlocked for everyone

Starting today, every Community Tier organization can:

Create your own repos. Build version-controlled, cloud-optimized data cubes with Icechunk, backed by Arraylake’s catalog.
Get 10 GB of managed storage. Every org comes with a Cloudflare R2 bucket that Earthmover provisions and manages for you.
Connect your own object store. Bring your own S3, GCS, R2, Azure, or any S3-compatible object store bucket to retain full ownership and sovereignty over your data. Arraylake manages the catalog while you keep the bytes.
Serve data with 50 compute credits per month. Spin up Earthmover Compute services (EDR, OPeNDAP, Tiles, WMS, openEO, and soon SQL) on top of your repos.
Invite members to your org. Collaborate with your team on shared repositories.
Create API clients and tokens. Wire up automated ingestion jobs, QA pipelines, and services with tightly-scoped credentials.
Create webhooks. Trigger downstream workflows whenever a repo is updated. You can easily rebuild a derived product, refresh a dashboard, or notify a channel.
Subscribe to Marketplace datasets. Pull in as many free and open datasets from the Data Marketplace as you like.

Up to 10 GB of managed storage, on us

For many scientists and developers, the hardest part of going cloud-native isn’t the tooling, it’s getting access to the cloud in the first place. Standing up a cloud account means a credit card, IAM policies, and credential management. That’s a lot of friction to give cloud storage a try.

The Community Tier of Arraylake removes that barrier entirely. In addition to being able to bring your own bucket (BYOB), every Arraylake organization, free or paid, now includes 10 GB of managed storage, so you can start writing data the moment you sign up. There are no cloud accounts to configure or credentials to manage. Simply click a button to provision your new managed storage bucket and start creating Icechunk datasets and committing data. Within minutes, you’re working with versioned, cloud-optimized arrays in real cloud object storage, even if you’ve never opened a cloud console in your life.

This managed storage is meant to act as a sandbox, providing a completely frictionless way to evaluate the platform, test out our data delivery APIs, or build a proof-of-concept before hooking up your own infrastructure.

Managed compute and configurable services

Data storage is only half the story. The Community Tier also unlocks Earthmover Compute, our managed API layer for data delivery. Every organization now gets 50 compute credits to run Compute services against their repos, unlocking:

SQL: (Coming soon!) Connect a PostgreSQL or Flight SQL client and query your data cubes with a subset of SQL.
openEO: Run Earth observation processing workflows against your data with standard openEO clients.
EDR: Point, area, and trajectory queries against your data cubes, returning CSV, CoverageJSON, GeoJSON, NetCDF, Parquet, or even TIFF
Tiles and WMS: Put your data on a slippy map in a browser or GIS application.
OPeNDAP: Remotely slice into your data cubes from the huge ecosystem of DAP-compatible tools in nearly any programming language.

To ensure your credits aren’t wasted, services on the free tier automatically spin down after 15 minutes of inactivity.

Alongside the free credits, we’ve also shipped a set of compute configuration improvements that benefit everyone. You can now pick the instance size for your Compute services (from XS up to 3XL) and number of instances to match your exact workload. Dial things down for a hobby project sipping credits, or scale up for a production service handling real traffic. Scale horizontally or vertically with your workloads. Credit usage scales with instance size and count.

Here is a look at what 50 credits gets you across different instance sizes:

On-Demand Maps (XS Instance): Host a Tile Service with up to 50 hours of uptime, which can stretch across several weeks of moderate sporadic web traffic.
Daily SQL Analyses (M Instance): Get over 12 hours of balanced compute, giving you about a half an hour every weekday for an entire month to run complex SQL queries.
Heavy EDR Extractions (XL Instance): Tackle complex EDR queries on a multi-terabyte dataset with over 3 hours of high-performance compute and enough memory to hold large data chunks for highly efficient processing.

Why we’re doing this

Earthmover was started because working with scientific data was too hard and too slow. Everything we’ve built since—Arraylake, Icechunk, Earthmover Compute, and the Marketplace—has been in service of making scientific data accessible. To make that mission a reality, we want to ensure the barrier to entry is as low as possible.

The Community Tier means anyone can derive real value from the full Arraylake platform: ingest and manage your data, serve it through standard APIs, and build your projects on top of it, all for free. Lowering the barrier also means lowering it to the cloud itself: the Community Tier gives you cloud object storage out of the box, with no third party cloud accounts to set up and no credential management required.

As the name implies, we are opening up Arraylake for the community, and we have built it to last. We know that relying on a free tier requires trust; you need to know the rug won’t be pulled out from under you in six months. Because of how we have engineered Arraylake, offering this Community Tier is completely sustainable for our business. This isn’t a temporary promotional phase; it is a dependable, permanent foundation that you can confidently build your workflows on.

And when your project grows into something bigger, the platform grows with you. Whether you need unlimited storage, more compute, or enterprise access controls, our team is here to help you find the right plan. The best part is that you can scale up without ever migrating your data or rewriting your code.

Get started

The refreshed Community Tier is live today for all users. If you have an existing free organization, it has already been upgraded, with nothing to migrate. If you don’t have an account yet, it’s easy to get started: