Engineering blog
Practical notes on building crypto backends — grounded in the same primitives the API exposes, no fluff.
Archive nodes, explained: what you actually pay for
Why archive nodes cost so much more than full nodes, what historical state unlocks, and how to tell whether an RPC provider truly serves archive or quietly prunes it.
Readtrace_ vs debug_: which tracing namespace to reach for
EVM nodes expose two overlapping tracing namespaces. Here is what trace_ and debug_ each do best, where they differ across clients, and how to pick one for your use case.
ReadReorgs and confirmations: how many blocks is safe?
Chain reorganizations are normal, not exceptional. Here is how deep they realistically go across networks and how to pick a confirmation threshold that fits your risk.
Readwei, gwei, ether: getting units right in production
Off-by-a-few-decimals unit bugs are the most common and most expensive class of error in crypto backends. A precise walk through the ladder and how to avoid float.
ReadConsolidating the crypto backend stack behind one key
Most teams stitch together an RPC provider, an indexer, an AML vendor and a custody API. The integration tax of four keys, four SLAs and four data models is underrated.
Read