​

Install Lasso in your app

# .env
LASSO_RPC_URL=https://lasso.sh/rpc/k/lasso_…
LASSO_WS_URL=wss://lasso.sh/ws/rpc/k/lasso_…

​

viem

import { createPublicClient, http, webSocket } from "viem";
import { base } from "viem/chains";

const client = createPublicClient({
  chain: base,
  transport: http(`${process.env.LASSO_RPC_URL}/latency-weighted/base`),
});

const subscriber = createPublicClient({
  chain: base,
  transport: webSocket(`${process.env.LASSO_WS_URL}/base`),
});

​

ethers v6

import { JsonRpcProvider, WebSocketProvider } from "ethers";

const provider = new JsonRpcProvider(
  `${process.env.LASSO_RPC_URL}/latency-weighted/base`,
  8453,
  { staticNetwork: true, polling: true },
);

const subscriber = new WebSocketProvider(`${process.env.LASSO_WS_URL}/base`, 8453);

​

Foundry, Hardhat and indexers

cast block-number --rpc-url "$LASSO_RPC_URL/base"
forge script Deploy --rpc-url "$LASSO_RPC_URL/fastest/base" --broadcast

// hardhat.config.ts
networks: {
  base: { url: `${process.env.LASSO_RPC_URL}/fastest/base` },
},

​

Choose the URL per call site

​

Roll out one path at a time

1. Pick one stateless read path your app already measures, such as a balance or eth_blockNumber poller.
2. Point it at Lasso with an explicit strategy, and keep the previous URL as a rollback.
3. Compare correctness, error rate, latency and head freshness against the old endpoint, and check routing evidence to see which provider served each request.
4. Move historical reads, transaction sending and subscriptions as separate steps; each has its own behavior in RPC method behavior.
5. Expand traffic once your own evidence supports it.

​

Limits

• Lasso rejects HTTP filter methods (eth_newFilter and friends). Use eth_getLogs or WebSocket subscriptions.
• A signed transaction is sent to one provider and not retried after dispatch. If the response is lost, look up the transaction hash before resending.
• JSON-RPC batches are routed per item; a batch is not pinned to one provider or one block.

​

Next

• Keys, URLs and profiles: defaults, scoping and rotation.
• Routing: what each strategy does.
• Multi-provider failover: test the failure you need to survive.