Sui's own GraphQL server, run against our unpruned indexers. Nested reads in a single round trip, cursor pagination, introspection left on, and the same fields answering for genesis as for the chain tip.
query {checkpoints(last: 3) { nodes {sequenceNumbernetworkTotalTransactions} }}{ "data": { "checkpoints": { "nodes": [{ "sequenceNumber": "98241733", ...
A view that wants balances, owned objects and recent transactions is one round trip. Credits follow what the query resolves, one per 4 output nodes, so a lean selection costs less than a greedy one and you never pay per call.
Cursor pagination runs over the same store that holds genesis, so paging back through a year of transfers is the same query shape as paging back through an hour.
simulateTransaction returns the effects without touching the chain. Both it and the execute mutation cost 5 credits, so checking first is never the expensive option.
The endpoint runs Sui's own graphql-rpc server against our unpruned indexers, so the schema is theirs and the history is ours. Reads are metered by what they resolve: one credit per 4 output nodes, live or archival alike.
An unparseable query is charged the base rate.See what every call costs on the pricing page ->
Credits track output nodes, one per 4, so a query is billed by what it returns rather than a flat fee per call. Select less, pay less; the deep screen still beats a pile of separate reads on round trips.
Archival data answers through the ordinary schema. There is no separate historical endpoint and no archive surcharge, so a query for 2023 looks like a query for last block.
Codegen, IDEs and devtools read the schema straight off the endpoint. Nothing is vendored into your repo where it can quietly fall behind a Sui release.
Rate limits and credit quotas are separate. Compare every plan ->
Free tier, no card required. The same key also answers on gRPC and JSON-RPC.