Blockchain RPC. One key for 35+ chains, flat request-unit pricing, failover you never notice. In production since 2021. ISO 27001 certified.

Dubai
Joined September 2024
Gateway round-trip on our networks page, measured on 18 Oct: Starknet 9 ms, Aptos 10 ms, Base 13 ms, Ethereum 14 ms, Solana 29 ms. Our monitors call every endpoint about once a second, and the page shows the live figure, not an average. Measured live, not a marketing number.
1
Hoodi's Glamsterdam fork is on hold, but Geth v1.17.9 and Lighthouse v8.3.x still schedule it for 26 Oct. Running your own Hoodi node? Check your execution and consensus clients for releases that drop the fork, or it can fork off at 17:42 UTC.
If you run Nethermind on Hoodi, upgrade to 2.2.1 before 26 Oct. On today’s All Core Devs call, client teams postponed the Glamsterdam fork on Hoodi to review a change to its specification. A new date will follow. Meanwhile, upgrade your Hoodi node now: github.com/NethermindEth/net…
2
Celo update: op-node celo-v2.2.4 and op-reth celo-v1.0.7 are live on our Celo endpoints. The op-node release from @CeloDevs fixes a deadlock in p2p peer discovery that could leave a node waiting indefinitely. No changes needed on your side.
7
Denim from @buildonbase reaches Base Sepolia today. Mainnet is planned for November. Base goes from one canonical block every two seconds to five per second, and Denim replaces Flashblocks. Our mainnet endpoints run base/base v1.5.0. Reading Flashblocks? Plan the move now.
18
Alpenglow from @anza_xyz activates in epoch 1064, expected Wed 21 Oct. Our Solana endpoints run Agave 4.3.x. Finality drops to a target of about 150 ms, and confirmed and finalized become one state. Commitment levels keep working on the day. Then use finalized.
25
If your app pipelines WebSocket requests to Nethermind nodes, match responses by id. On 2.2.0 they can come back out of order.
Nethermind Client 2.2.0 is out. For RPC providers: faster responses that use less memory, and faster tracing on Glamsterdam blocks. Mandatory for Hoodi nodes before the Glamsterdam fork on Oct 26. github.com/NethermindEth/net…
1
25
Chain uptime and your app's uptime are two different things. A healthy chain can't help if your app reaches it through one RPC endpoint. Test your failover before the day you need it.
1
36
Sonic v2.2.5 from @SonicLabs is live on our endpoints. Custom JavaScript tracers are now off by default, and we keep them off. debug_trace* with a custom JS tracer fails with "custom tracer is not permitted". Using one? Built-in tracers work as before.
37
Geth v1.17.8 from @go_ethereum is a security release. It closes several DoS risks; most fixes cover bugs only possible after Amsterdam. It also fixes eth_simulateV1 transfer logs and eth_createAccessList for senders with no balance. Run your own Geth node? Upgrade.
15
eth_chainId round-trip from our gateway, measured on 11 Oct: Ethereum 27 ms, Base 27 ms, Arbitrum 26 ms, Optimism 26 ms, Polygon 27 ms, BNB Chain 28 ms. Measured live on the homepage. Check today's figures yourself.
1
17
Sui's testnet-v1.82.1 already drops JSON-RPC from full nodes. sui_* and suix_* calls must move. Sui gRPC (sui.rpc.v2) is in beta on your BoltRPC key. Pick fields: on our endpoint (12 Oct), a checkpoint was 76 bytes with a minimal read mask, about 160 KB with full transactions.
1
19
Starknet L1 to L2 message fee estimates now have a floor. Juno v0.16.8 from @Nethermind is live on our endpoints. starknet_estimateMessageFee returns overall_fee = max(5 × 10^13 wei, estimate), the minimum L1 already enforces. Summing fee fields yourself? Use overall_fee.
1
1
39
Polygon's Lugano hardfork went live on 1 Oct at Heimdall block 54,627,000. Our Polygon endpoints ran through it without interruption. Heimdall v0.12.1 from @0xPolygon carries Lugano's security fixes. Nodes on v0.11.0 or older had to upgrade. No action needed on your side.
39
Keys leak: a repo, a log, a screenshot. Scope each one so a leak can only do what you allowed. Lock each BoltRPC key to your origins or IPs, its chains and its methods, with its own rate limit and daily cap. Roll a key and the old one works up to 7 days. Deploys don't break.
1
25
A node that fails, falls behind or stops advancing is taken out of rotation. Traffic moves to a healthy one, no code changes on your side. WebSocket subscriptions stay open, and blocks missed during the switch are replayed. Failover handled for you, on every plan.
1
20