Engineering
Design trade-offs and architecture for reading blockchain data in production.
Engineering posts are about building on blockchain data in production: which transport to stream with, how to page through history without tripping range limits, how to size a plan against per-second rate limits, and how to recover when a stream drops. Each post compares behaviour across the networks we serve, works through the arithmetic with published limits, names the cases where the approach is the wrong one, and ends with a dated list of sources. For step-by-step integration, follow the links into the docs; for protocol concepts on their own, see Learn. Docs · Learn
Alpenglow is live on Solana mainnet: what block consumers and dashboards need to change
Anza says finality dropped from 12.8 seconds to about 150 milliseconds. The claim is vendor-reported; the changes to votes, commitment and block streams are what your code will actually meet.
The BNB Smart Chain methods the Triport dedicated pool now covers
BSC used to be the read-only corner of the gateway. A dedicated pool closes the methods that a public gateway could not — wide log ranges, full history, debug traces and a WebSocket — and this is where each one lands against what other providers publish.
Solana RPC latency: what we broke, what we fixed, and how many milliseconds our gateway adds
For six days a routing rule sent some recent-block reads to a slow node. This is the postmortem, the fix, the before/after numbers per network, and the phase histogram we now use to show how much time our own code adds to each request.
How a Solana launchpad indexer turns validator events into token cards
A launch feed is a sequence of observations, not a list of guaranteed opportunities. Follow the path from validator updates to a token card and learn what each stage can tell you.
Pump.fun bonding curves, almost bonded tokens and migration explained
Curve progress measures a launch stage. Migration changes the venue. Neither tells you whether a token is safe or what your sell will receive. Read the stages and card signals together.
Zero platform fee Solana trading — what you still pay
A zero platform fee removes the terminal's charge. Protocol fees, transaction costs and execution price still determine what reaches your wallet. Here is how to separate them.
Racing multiple block feeds: how we deliver the earliest copy of every Robinhood block
A single upstream feed makes you wait for whatever one source does next. We now run several in parallel, forward each block from whichever arrives first, and keep the rest as live fallback.
Base and Robinhood Chain side by side: what a trading bot must change
Both chains are EVM L2s that pay gas in ETH, so a Base bot compiles and connects to Robinhood Chain on the first try. Then fees, ordering, confirmation and the early-data stream all behave differently. A row-by-row map of what to rewrite, with live answers from both chains.
First-come, first-served: pricing transactions on Robinhood Chain
The EIP-1559 fields are all there, wallets fill them in, and none of them buys position. What the fee methods actually returned on 2026-10-13, where the L1 data fee hides, and a submission policy for bots on a first-come, first-served sequencer.
What changes when you port a dApp from Ethereum to Robinhood Chain
Robinhood Chain runs Solidity unmodified, which is exactly why the differences hurt — they compile, deploy and pass local tests. Nine places where Ethereum assumptions go quiet, each with the RPC call that shows the difference on the live network.
Reading Robinhood Stock Tokens on-chain: balances, uiMultiplier and transfers
A Stock Token balance is not a share count, and the price feed already knows that. Which contract values to read over JSON-RPC, the one formula that connects them — checked against the chain to the last unit — and the transfer events that are easy to count twice.
eth_getLogs block range limits on Ethereum, BSC, Base, Polygon, TRON and Robinhood Chain
The block-range ceiling on eth_getLogs is not one number per chain — it can differ between nodes behind the same URL. What the six EVM-style chains we serve document, and a pager that survives all of them.
Sizing an RPC plan from your workload: sustained RPS, burst and heavy-method buckets
A request-per-second plan is sized differently from a credit plan. How to measure a workload, how a token bucket with 2× burst actually behaves, and a worked Solana indexer that lands on a specific tier.
SSE vs WebSocket vs gRPC: choosing a transport for blockchain data
Three ways to receive a push from a blockchain node, compared on what they actually guarantee — direction, framing, auth, resume and filtering — with one real chain for each.
Watching incoming payments on Solana, Ethereum, Stellar and TON: one task, four designs
Detecting a deposit looks like the same feature on every chain. It is four different designs: an account subscription, a log subscription with a blind spot, a stream that only says when to look, and a poll.
Yellowstone gRPC filters in practice, and the same filter set over a WebSocket bridge
A Yellowstone subscription is one request describing everything you want. How its fields combine, what happens when you change it, and how to split a workload across a fixed number of streams.