Every method on BoltRPC has a published request-unit weight. eth_call is 3 RU, eth_getLogs is 4, debug_traceTransaction is 6. That is the whole pricing model. Multiply your calls by their weights and you know the bill before you ship. boltrpc.io/pricing

Get the fastest 33 public endpoints for development on Ethereum mainnet from your location and hosted by @tatum_io @getblockio @1rpc_ @Alchemy @Allnodes @blink_labs_xyz @blockmachine_io @BlockRazor_Inc @blockreq @bloxroute @CoWSwap @dRPCorg @Nodeflarerpc @nodies_infra @NOWNodes @OnFinality @POKTnetwork @polka_chu @RealBlockPI @reddio_com @SenseiNode @sentioxyz @solidrpc @Stakely_io @TenderlyApp and @thirdweb Try them in our library at comparenodes.com/library/pub…

Next in our synthetic WebSocket series: the @ethereum Beacon chain. The Beacon chain is the consensus layer, and it speaks REST rather than JSON-RPC. That keeps it clean, but it also means there is no subscription API to connect to. So if you track consensus, you poll every twelve seconds and hope your timing lines up with the slot. We stream the slots instead. Each one reaches you as it lands at the head, with the slot number, block root and proposer, pushed from our own proxy. Empty slots are handled properly. When no block is proposed, which is entirely normal, the stream does not report a gap or invent one. After a failover, missed slots are backfilled in order. And when you need consensus events rather than heads, we relay the standard event stream over HTTPS, so your existing tooling keeps working. Same on the mainnet Beacon chain plus the Hoodi and Sepolia beacon testnets. Stay tuned for the next chain in this series.

Fourth chain in our synthetic WebSocket series: @trondao. Tron settles the largest share of USDT transfer volume, which means most of what moves across it is payments. And payments only ever ask one question: is this final? Tron nodes do not carry a subscription API so the usual answer is to poll for new blocks. We push them to you instead, streamed from our own proxy the moment they land. For the finality question we serve Tron's solidified view: the same reads, answered only from state that can no longer be reversed. You ask, you get an answer you can act on. Both sides of Tron run on one endpoint. The native wallet API that TronWeb teams already use, plus an Ethereum-style JSON-RPC surface so existing EVM tooling points at Tron with almost no code change.

Third chain in our synthetic WebSocket series: @SuiNetwork. Sui produces checkpoints faster than most infrastructure can keep up with. That is a compliment to the chain, and a genuine engineering problem for anyone reading it. Sui nodes do not offer a WebSocket, so polling is the default answer. On a chain this fast, polling means you learn what changed after the fact and miss whatever happened in between. So we push instead. The latest checkpoint reaches you the moment it lands, streamed from our own proxy. The full Sui read surface is there for your queries, without the unsafe methods that build transactions for you server side. Your keys stay yours. Behind it sit multiple Sui nodes with automatic failover. If one drops, your stream keeps running.

Synthetic WebSocket series: @Bitcoin Polling Bitcoin once a second costs you 2.6 million requests a month. Streaming it costs about two cents. Bitcoin is live on BoltRPC. Full node access over HTTPS, plus a live block stream we synthesize ourselves, because Bitcoin nodes do not offer one. wss://endpoints.boltrpc.io/ws/KEY/bitcoin When Bitcoin reorganises and a block gets replaced, we send you the new one straight away. You are never left holding data that is no longer true. And you only need one connection. The same socket that streams your blocks also handles your normal Bitcoin calls. Read-only by design. Wallet, key-handling and node-control RPCs are never relayed. That is a guarantee, not a gap. Same key, same dashboard.

Introducing synthetic WebSocket support. We crossed 35 networks, as many of you noticed. Here is the part that matters more: the chains that have no native WebSocket now get one from us. @Aptos produces blocks in under a second, which is a big part of why it feels instant to build on. It is also built around a REST API, which keeps the node lean. What it does not carry is a subscription API, so teams that want live updates end up polling for them. This changes now. BoltRPC synthesizes the block stream inside our own proxy, from the same node monitoring that drives our failover. To your client it looks exactly like a native WebSocket chain: wss://endpoints.boltrpc.io/ws/KEY/aptos eth_subscribe works, so your tooling runs unmodified. Every Aptos block header carries a blocks_skipped count, so you always know what you skipped. Aptos is live on BoltRPC. Stay tuned for the other chains with no native WebSocket. We are posting them one by one. On boltrpc.io you can see every network we serve. The ones with WebSocket are marked WS.

The new BoltRPC dashboard is live. Your keys, your usage, your logs and your team, all in one place. What you get the moment you log in: Team access. Every member of your team gets their own keys and their own usage view, while the owner manages permissions plus billing in one place. Request-level logs. Not a summary graph. The actual requests, filterable by key, network and method, with export. Per-method pricing, published upfront. You can see what every call costs before you build against it, on a dedicated Method pricing page. No guessing what a request will bill. Usage and billing in the same view. Spending limits, usage alerts, an on-demand overview that matches what you are actually charged. An RPC playground per network, so you can fire a call and see the response without leaving the browser. We built this because the feedback was consistent: teams operating at scale do not want a support ticket to find out what happened. They want their own logs and their own numbers. It is live now at dashboard.boltrpc.io.

A faulty price feed on @Starknet lending protocol Vesu triggered 3 million dollars in liquidations on September 4. Oracle provider Pragma published a wrong USDT/USD price for about 30 minutes. Vesu made 47 liquidations across 42 wallets, all within two minutes.

60s onboarding is a win

Hanwha, a 200 billion dollar Korean conglomerate, launched a tokenization platform on @avax on September 7 to settle tokenized securities. It lands alongside South Korea's phased framework for blockchain securities, effective February 2027. boltrpc.io/networks/avalanch…

OP Stack update: op-reth v2.4.4 and op-node v1.19.7 are now running on our OP Stack Chains. op-reth restores canonical-head fee and gas metrics, on par with op-geth. op-node is a build-hygiene release only, no consensus change.

Robinhood Chain is live on BoltRPC. @RobinhoodCrypto's Arbitrum L2 for tokenized stocks, built for 24/7 trading. Settles to Ethereum via blob data availability, ETH for gas, ~100 ms preconfirmations. Uniswap and Chainlink live from day one. boltrpc.io

#BOB @build_on_bob is now in Top 75 protocols in our directory for web3 infrastructure with new support by @BoltRPC Get your public or private RPC endpoints for BOB at comparenodes.com/protocols/b…

We reworked boltrpc.io with a cleaner UX, live chain latency, and pricing you can actually read before you ship. Same multi-chain RPC (24 chains, one key). New site to explore it all. Take a look → boltrpc.io

Live since 2021, trusted by teams like Chainlink, Enjin, Tiingo and Gains Network. 20+ chains, one API key. Two weeks free, no credit card. boltrpc.io

Starknet Juno Update ! The Starknet ecosystem continues moving fast and the latest updates to Juno, Nethermind’s Go-based Starknet full-node client are a great example of this progress. 🔧 What’s New in Juno’s Latest Release 🔧 This release introduces a new database architecture that cuts Starknet node storage from 500GB down to 200GB when paired with modern compression. Juno v0.15.18 also opens up new performance tuning controls, letting operators configure their nodes based on actual hardware, instead of one-size-fits-all defaults. With zstd compression now available (and becoming default in 0.16.0), database growth slows dramatically without CPU trade-offs. It’s an optional migration for now, but it won’t stay that way. This release reflects ongoing refinement at Starknet’s infrastructure layer. Focusing on long-term cost efficiency, performance consistency and sustainable production operations.