Independent Blockchain Infrastructure

We run RPC nodes on dedicated capacity from multiple providers, across multiple regions. When one provider fails, the others often don't. Our edge is decorrelation.

AGENT CORE US-WEST US-CENTRAL US-EAST EU-WEST EU-NORTHWEST ASIA-SOUTHEAST ASIA-NORTHEAST
— Dedicated hosts updated —
— Regions
— Mainnet chains in production —

View ethereum-rpc-docker

How it runs itself

Sense fleet telemetry · market data · alerts SENSORS + ETL Store one warehouse of queryable facts DATA WAREHOUSE Model predictors trained on retained history STATE + MODELS Decide policy maps state to action MANAGER + CONTROL LAW Act tracked, idempotent workflows MECHANICS + JOB ENGINE Learn every outcome graded nightly GRADERS + PROVENANCE THE LOOP

The operating loop — sense, store, model, decide, act, learn — is the engine of the company. The feedback edge that closes the loop is what makes it autonomous. Read the build log →

What We Do

Dedicated Capacity

Dedicated servers from multiple providers. No shared tenancy, no noisy neighbors. Predictable performance.

Multiple Regions

loading...
Geographic distribution means failures stay local.

In Production

Ethereum and the L2 ecosystem, plus a long tail few others serve. Right now the fleet runs — mainnet chains (— testnets). The product ships more supported configurations than we run.

Our Philosophy

Independence

We run the whole ops stack: client software, upgrades, monitoring, regional placement. Multiple providers and regions keep our failures uncorrelated with any one hyperscaler.

Measurement

Every node, every chain, every request is measured. We know what's working and what isn't, in real time.

Leanness

We're building toward a zero-human company — agents run the show, humans set direction.

ethereum-rpc-docker

We open-source the docker-compose configurations we use in production — supported configurations for 90 mainnet chains (+81 testnets). On the managed fleet we run a subset in production, kept in sync, upgraded through hardforks, monitored 24/7.

You can run this yourself — it's free, documented, and we help. But there's an easier way: buy the managed service from us. We run it on our fleet, we handle the upgrades, you get reliable RPC.

View on Gitea

Network client diversity

A live census of execution and consensus clients (and versions) observed in the peer sets of nodes we run — aggregates only, no peer identities.

View the census

Need a Snapshot?

Many long-tail blockchain networks have no public snapshots. Syncing from genesis can take days to weeks. We sell the exact snapshots our production fleet uses — fresh, verified, with client version and head block listed.

$90 per TB — anchored to cloud egress pricing. Launch offer: your first snapshot is free, just email us.

Browse Snapshots

Inspiration

This project is inspired by Brian Roemmele's "5000 Days" podcast series, exploring what's possible when we rethink how companies operate.

Brian Roemmele on X Read Multiplex