Keep your RPC provider. Cut your bill.

Edge Boost is a cache and global CDN that sits between your app and your RPC provider. When your app asks for blockchain data Goldsky has already indexed, Boost answers directly in milliseconds. Everything else is forwarded to your provider exactly as you sent it, and the response comes back exactly as they answered it. You keep your provider, your credentials, and your contracts. Boost just answers the repeat questions before they reach your provider.

There isn't one. Goldsky already populates and stores this chain data to power Turbo and other products, so the expensive work is a fixed cost we already carry. Querying that database even 100 times more costs little compared with indexing and storing it in the first place. Instead of every team privately repeating the same RPC calls, Boost opens the data Goldsky already maintains for anyone to use. We are already doing the work; Boost lets the rest of the industry share it.

No, and that's the point. Boost works in front of whichever provider you use today. You tell us where to forward requests, swap one URL in your app, and everything else stays the same. If you ever want out, swap the URL back and you're done.

All of them, as long as the endpoint is reachable over https. Alchemy, QuickNode, Infura, dRPC, Nirvana Labs, Ankr, Chainstack, a node you run yourself — if your app can call it, Boost can front it. Providers that authenticate with a bearer token or API-key header work too: custom headers are configured once and applied to every forwarded request. You can even put Boost in front of Goldsky's own Edge RPC.

One URL. Point your RPC client at your Boost endpoint instead of your provider's, and you're running. Boost speaks standard JSON-RPC, so viem, ethers, foundry, and every other client work unchanged.

Anything time-sensitive, like latest, pending, safe, or finalized reads, always skips the cache and goes straight to your provider. Cached answers come from Goldsky's indexed blockchain data, which follows your source chain and picks up reorgs within about 100 milliseconds. And every response includes a header that says whether it came from the cache or your provider, so you can verify us from your very first request.

The request forwards to your provider byte-for-byte, and the response comes back untouched. Forwarded requests are free. A miss costs a little latency, never correctness.

Yes, and nothing to enable: post an array to the same address. Boost serves a batch per item, so the calls it holds are answered from cache and the rest are forwarded to your endpoint in a single onward batch. Give every item a unique id, because JSON-RPC makes no promise about the order of a batch response, so match results by id rather than by position. Arrays over 100 items are forwarded whole, verbatim, with no cache lookups, so split larger arrays into 100-item batches to keep the hits.

x-edge-billable carries the count of items Boost served from cache. x-cache reports HIT only when every item in the batch came from cache, so a batch where 44 of 45 items were cached still reports MISS. Read the served count rather than x-cache alone.

At launch: Ethereum, Polygon, Base, Arbitrum, Optimism, BNB Chain, HyperEVM, and Robinhood, with more on the way. Boost caches historical reads like eth_getBlockByNumber, eth_getTransactionReceipt, eth_getTransactionByHash, eth_getLogs, and eth_getBlockReceipts. Everything else forwards to your provider. WebSocket connections aren't supported at launch, so keep those pointed at your provider.

Edge RPC is Goldsky's fully managed RPC service: we run the endpoints, routing, and failover for you. Edge Boost keeps your current provider in place and adds a cache in front of it. They work well together, too: you can use Edge RPC as the provider behind Boost.

Yes. Your provider URL and any headers you configure are encrypted at rest, and you can replace or remove them in the dashboard at any time. You can also keep standby providers on file per chain and switch the active one whenever you like.

No. Goldsky populates Boost's cache from our own internal indexing workloads and the chain-level indexing that powers Turbo and other Goldsky products. Boost customer requests and response data are not used to build the cache.

Edge Boost is completely free: there are no usage fees for cache hits or requests forwarded to your provider. Enterprise is available for teams that need committed support response times, uptime SLAs, custom configurations, private networking, or other advanced functionality.