Every service Sui exposes, served from live nodes and full archival storage since genesis. One API key, no pruning horizon, no second provider for history.
$ grpcurl -H 'x-api-key: indr_pat_qsVO...' \mainnet-grpc.inodra.com:443 \sui.rpc.v2.LedgerService/GetServiceInfo{ "chainId": "35834a8a...","checkpointHeight": "98241733","lowestAvailableCheckpoint": "0" }
Batch reads keep the round trips down while credits follow the entities returned, one each, doubled when served from the archive. The bill tracks how much history you actually pull, never how you shape the calls.
Every checkpoint arrives as it commits. Streaming is metered per 50 KB, so a field mask that drops what you never read cuts the bill directly.
Get the effects of a transaction without touching the chain, then submit it. Both calls cost 5 credits, so a dry run is priced exactly like the send.
The proxy exposes Sui's native gRPC surface unchanged - the schema is Sui's own, so codegen runs off Sui's published protos and nothing Inodra-specific needs vendoring. Reads are metered per entity returned: a 200-key batch is 200 credits, and archival-served reads cost double.
Reads served from the archival tier cost 2x. Unlisted methods fall back to 1 credit.See what every call costs on the pricing page ->
Ask for a checkpoint from last year and it returns. The same call against a pruned node fails, and pruning is never something you find out about at a convenient moment.
One address decides whether a request is served from a live node or from archival storage. Your client never changes.
A read served from the archive costs double, keyed on where the gateway routed it, never on whether a cache happened to be warm. A deep backfill's bill is arithmetic, not a surprise.
Rate limits and credit quotas are separate. Compare every plan ->
Free tier, no card required. Archival history is included from the first key.