// use cases
Where state bloat hits the invoice.
The library earns its keep wherever a growing chain turns storage and data movement into the bill that matters and the data is Solana-shaped — archive nodes, DePIN storage networks, indexers, explorers. Four workloads where that is true, each with the preset it aims at.
Archive nodes
up to 60:1// Full history without a cost curve that tracks block height
A growing chain adds state forever, and archive-node operators pay to keep all of it online. Cold, write-once history is exactly what MaxCompression is for — the measured "up to 60:1 on archival workloads," and CPU cost you only pay once at write time.
preset: MaxCompression →
Account state snapshots
15–40:1// Snapshot dumps compress hardest of all
Account snapshots are dense with repeated owner fields, program IDs, and key shapes. The Accounts preset targets exactly this structure and reports the highest non-archival ratios in the README.
preset: Accounts →
Block explorers
12–35:1// Keep more of history on warm disk
As history piles up, explorers serve an ever-longer tail. Smaller stored size means more blocks stay on warm storage and fewer queries pay the cold-tier round-trip. Mixed is a safe default across transactions and account state, and its illustrative range is workload-dependent.
preset: Mixed →
Indexer & DePIN warehousing
10–30:1// Warehouse and replicate the firehose without the storage tax
Indexers and decentralized data networks warehouse decoded transactions — mostly repeated program IDs with a small tail of distinct arguments. The Transactions preset collapses the repeated part to a reference at level 3, so fewer bytes hit disk and fewer bytes cross the wire between replicating peers.
preset: Transactions →