➡️Writing Typescript Functions

1) Clone the Web3 Function Hardhat Template

git clone https://github.com/gelatodigital/web3-functions-hardhat-template.git
cd web3-functions-hardhat-template

or use the template on Github's UI.

2) Install dependencies

yarn install

3) Copy the example .env file

cp .env.example .env

4) Fill in your Alchemy & private key for local testing

Note: PRIVATE_KEY is optional, and only needed if you wish to deploy contracts or create a task from the CLI instead of the UI

ALCHEMY_ID= // <= Your Alchemy key
PRIVATE_KEY= // <= Your Private key

Typescript Function Example

This typescript function updates an oracle smart contract with data returned by Coingecko's price API at an interval. Check out more examples here.

import { Web3Function, Web3FunctionContext } from "@gelatonetwork/web3-functions-sdk";
import { Contract, ethers } from "ethers";
import ky from "ky"; // we recommend using ky as axios doesn't support fetch by default

const ORACLE_ABI = [
  "function lastUpdated() external view returns(uint256)",
  "function updatePrice(uint256)",

Web3Function.onRun(async (context: Web3FunctionContext) => {
  const { userArgs, gelatoArgs, multiChainProvider } = context;
  const provider = multiChainProvider.default();

  // Retrieve Last oracle update time
  const oracleAddress = "0x71b9b0f6c999cbbb0fef9c92b80d54e4973214da";
  const oracle = new Contract(oracleAddress, ORACLE_ABI, provider);
  const lastUpdated = parseInt(await oracle.lastUpdated());
  console.log(`Last oracle update: ${lastUpdated}`);

  // Check if it's ready for a new update
  const nextUpdateTime = lastUpdated + 300; // 5 min
  const timestamp = (await provider.getBlock("latest")).timestamp;
  console.log(`Next oracle update: ${nextUpdateTime}`);
  if (timestamp < nextUpdateTime) {
    return { canExec: false, message: `Time not elapsed` };

  // Get current price on coingecko
  const currency = "ethereum";
  const priceData: any = await ky
      { timeout: 5_000, retry: 0 }
  price = Math.floor(priceData[currency].usd);
  console.log(`Updating price: ${price}`);

  // Return execution call data
  return {
    canExec: true,
    callData: [{
      to: oracleAddress,
      data: oracle.interface.encodeFunctionData("updatePrice", [price])

Create your function schema.json to specify your runtime configuration.

  "web3FunctionVersion": "2.0.0",
  "runtime": "js-1.0",
  "memory": 128, 
  "timeout": 30,
  "userArgs": {}

Note: For now the configuration is fixed and cannot be changed.

Typescript Function Context

When writing the Web3 Function, it is very helpful to understand the context Gelato injects into the execution, providing additional features to widen the Web3 Functions applicability.

Web3Function.onRun(async (context: Web3FunctionContext) => {
  const { userArgs, storage, secrets, multiChainProvider, gelatoArgs} = context;
  const provider = multiChainProvider.default();

User Arguments

  1. Declare your expected userArgs in your schema, accepted types are string, string[], number, number[], boolean, boolean[]:

  "web3FunctionVersion": "2.0.0",
  "runtime": "js-1.0",
  "memory": 128, 
  "timeout": 30,
  "userArgs": {
    "currency": "string",
    "oracle": "string"
  1. Access your userArgs from the Web3Function context:

Web3Function.onRun(async (context: Web3FunctionContext) => {
  const { userArgs, gelatoArgs, secrets } = context;

  // User args:
  console.log('Currency:', userArgs.currency)
  console.log('Oracle:', userArgs.oracle)
  1. In the same directory as your web3 function, create a file userArgs.json and fill in your userArgs to test your web3 function:

  "currency": "ethereum",
  "oracle": "0x71B9B0F6C999CBbB0FeF9c92B80D54e4973214da"

Test out the Coingecko oracle web3 function:

npx hardhat w3f-run oracle --logs

State / Storage

Web3Functions are stateless scripts, that will run in a new & empty memory context on every execution. If you need to manage some state variable, we provide a simple key/value store that you can access from your web3 function context.

See the above example to read & update values from your storage:

import {
} from "@gelatonetwork/web3-functions-sdk";

Web3Function.onRun(async (context: Web3FunctionContext) => {
  const { storage, multiChainProvider } = context;
  const provider = multiChainProvider.default();

  // Use storage to retrieve previous state (stored values are always string)
  const lastBlockStr = (await storage.get("lastBlockNumber")) ?? "0";
  const lastBlock = parseInt(lastBlockStr);
  console.log(`Last block: ${lastBlock}`);

  const newBlock = await provider.getBlockNumber();
  console.log(`New block: ${newBlock}`);
  if (newBlock > lastBlock) {
    // Update storage to persist your current state (values must be cast to string)
    await storage.set("lastBlockNumber", newBlock.toString());

  return {
    canExec: false,
    message: `Updated block number: ${newBlock.toString()}`,

To populate the storage values in your testing, in the same directory as your web3 function, create a file storage.json and fill in the storage values.

  "lastBlockNumber": "1000"

Test out the storage web3 function:

npx hardhat w3f-run storage --logs


  1. In the same directory as your web3 function, create a .env file and fill up your secrets.

  1. Access your secrets from the Web3Function context:

  // Get api from secrets
  const coingeckoApi = await context.secrets.get("COINGECKO_API");
  if (!coingeckoApi) {
    return { canExec: false, message: `COINGECKO_API not set in secrets` };
  1. Test your web3 function using secrets:

npx hardhat w3f-run secrets --logs
  1. When deploying a task, you will be able to set your web3 function secrets on our UI or using the SDK, see here

import hre from "hardhat";
import { AutomateSDK, Web3Function } from "@gelatonetwork/automate-sdk";

const { ethers, w3f } = hre;

const adBoardW3f = w3f.get("advertising-board");

const [deployer] = await ethers.getSigners();
const chainId = (await ethers.provider.getNetwork()).chainId;

const automate = new AutomateSDK(chainId, deployer);
const web3Function = new Web3Function(chainId, deployer);

// Deploy Web3Function on IPFS
console.log("Deploying Web3Function on IPFS...");
const cid = await adBoardW3f.deploy();
console.log(`Web3Function IPFS CID: ${cid}`);

// Create task using automate sdk
console.log("Creating automate task...");

const { taskId, tx } = await automate.createBatchExecTask({
  name: "Web3Function - Ad Board",
  web3FunctionHash: cid,
  web3FunctionArgs: {},

await tx.wait();
console.log(`Task created, taskId: ${taskId} (tx hash: ${tx.hash})`);
  `> https://beta.app.gelato.network/task/${taskId}?chainId=${chainId}`

// Set task specific secrets
const secrets = adBoardW3f.getSecrets();
if (Object.keys(secrets).length > 0) {
  await web3Function.secrets.set(secrets, taskId);
  console.log(`Secrets set`);

Multichain Provider

The multichainProvider allows us to instantiate RPC providers for every network Gelato is deployed on.

import {
} from "@gelatonetwork/web3-functions-sdk";

Web3Function.onRun(async (context: Web3FunctionContext) => {
  const { multiChainProvider } = context;

  // multichainProvider.default() will instantiate
  // the provider of the chain the W3F is deployed
  const provider = multiChainProvider.default();

  // passing the chainId as follows, we can instantiate
  // a rpc provider for that network
  const polygonProvider = multiChainProvider.chainId(137)

When testing locally, we can provide the different providers by including them in .env at the root folder.

// .env file

Interoperability with Other Libraries

Although multiChainProvider is designed to work seamlessly within the Gelato Web3 Functions SDK, it is possible to extract the underlying RPC URL and use it with other client libraries. This flexibility is valuable for developers who prefer or require features from other libraries, such as viem.

Here's an example of how to utilize the RPC URL from multiChainProvider with the viem library, which can be useful if you need to leverage features specific to viem:

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

Web3Function.onRun(async (context: Web3FunctionContext) => {
  const { multiChainProvider } = context;
  const provider = multiChainProvider.default();
  const url = provider.connection.url;

  // Initialize viem client with the extracted URL
  const rpc = createPublicClient({
    chain: polygon,
    transport: http(url),
  // Now you can use the viem client for your operations
  // ...

Gelato Arguments

Gelato injects the chainId, the gasPrice, and the taskId into the context.

  1. chainId: The unique number identifying the blockchain network where the function is running.

  2. gasPrice: The cost of executing transactions on the blockchain.

  3. taskId: A string that uniquely identifies the task.

import {
} from "@gelatonetwork/web3-functions-sdk";

Web3Function.onRun(async (context: Web3FunctionContext) => {
  const { gelatoArgs } = context;

  // chainId: number
  const chainId = gelatoArgs.chainId;

  // gasPrice: BigNumber
  const gasPrice = gelatoArgs.gasPrice;
  // taskId: string
  const taskId = gelatoArgs.taskId;

Last updated