Deep dives, guides and case studies from the team running supanode infrastructure. No fluff — just what actually works on Solana & beyond.
Hyperliquid CLOB: How HyperCore Matches Orders
How the HyperCore CLOB works across consensus batch ordering, order matching, L2 versus L4 data, latency, and production data access.
Jupiter API Integration: Custom RPC and Transaction Landing
Build a Jupiter Swap API V2 integration with custom Solana RPC, transaction simulation, priority fees, landing, confirmation, real-time monitoring, and historical swap data.
Pump AMM Data Streaming on Solana: gRPC vs WebSocket
Compare Yellowstone gRPC and WebSocket for streaming Pump.fun and PumpSwap activity, including lifecycle tracking, event decoding, account updates, reserves, backfill, and reconciliation.
Solana Data Streaming: WebSocket, gRPC, Shreds, and Indexers
A practical comparison of Solana WebSocket, Yellowstone gRPC, raw and decoded shreds, managed streams, and indexers by data stage, transport, and recovery needs.
Solana Historical Data: RPC, Indexers, and SQL Access
Compare archival RPC, historical APIs, self-built indexers, and managed SQL-first indexers for Solana historical data, analytics, and production queries.
How to Land Transactions on Solana: Fees, SWQoS, and Retries
Understand why Solana transactions fail to land, how priority fees, SWQoS, retries, and confirmation tracking work, and how to improve landing rate in production.
Solana ShredStream: Raw Shreds, Decoded Shreds, and Yellowstone gRPC
Solana ShredStream delivers raw shreds over UDP before RPC or gRPC, but the early signal only matters if decoding and recovery costs are justified.
Hyperliquid WebSocket API: Subscriptions, Limits, and Production Setup
A practical breakdown of Hyperliquid WebSocket subscriptions, public endpoint limits, reconnect behavior, recovery patterns, and production infrastructure choices.
Yellowstone gRPC on Solana: Geyser Streaming Guide
Learn how Yellowstone gRPC works on Solana, how it relates to Geyser, RPC, and WebSockets, and when production backends should use streaming instead of polling.