Step-by-step setup
Checkpoint is a Node.js library and running the following NPM commands will install it:
npm install @snapshot-labs/checkpoint
# or if using yarn
yarn add @snapshot-labs/checkpoint
Three arguments are required to initialize a Checkpoint instance. These are:
  • Checkpoint Configuration,
  • GraphQL entity Schemas,
  • Data Writers.
There is a fourth optional argument for configuring options like log levels and database connection. Read more about this parameter here.

Creating a Checkpoint Configuration

A checkpoint configuration is an object that defines the smart contract addresses and their respective events. For this guide, we will be indexing the list of posts and authors for this Poster Contract. A copy of this contract is deployed to the goerli-alpha network at the address: 0x04d10712e72b971262f5df09506bbdbdd7f729724030fa909e8c8e7ac2fd0012.
To successfully track the addresses of authors, you'll need to listen to new_post events from the contract. Therefore, a valid checkpoint configuration for the above requirements will be:
import { CheckpointConfig } from "@snapshot-labs/checkpoint";
const config: CheckpointConfig = {
network: "goerli-alpha",
sources: [
contract: "0x04d10712e72b971262f5df09506bbdbdd7f729724030fa909e8c8e7ac2fd0012",
start: 185778,
deploy_fn: "handleDeploy",
events: [
name: "new_post",
fn: "handleNewPost",
The start block number is set to 185778 because the contract was deployed at that block. This will mean Checkpoint starts scanning from that block as opposed to starting at block 0.
The deploy_fn and fn values are the names of the data writer functions to be invoked when the contract deployment and new_post events are encountered respectively. Read more about Checkpoint configuration here.

Defining GraphQL entity schemas

Checkpoint requires a set of defined GraphQL Schema Objects. These schema objects will be used to create the database tables for indexing records and also generate graphql queries for accessing the indexed data.
For this guide, we will want to track a Post entity and have it exposed via the graphql API. This entity can be defined as the following schema:
const schema = `
""" Entity named Post """
type Post {
id: String!
author: String!
created_at_block: Int!
Checkpoint will use the above entity (Post) to generate a MySQL database table named posts with columns matching the defined fields. It will also generate a list of graphql queries to enable querying indexed data. Read more about how queries are generated here.

Creating Data Writers

Data writers are Javascript functions that get invoked by Checkpoint when it discovers a block containing this event. A Data writer is responsible for writing records to the database. These records will eventually be exposed via Checkpoint's GraphQL endpoint.
We have defined two data writer functions in our Checkpoint configuration above. These are:
  • handleDeploy and,
  • handleNewPost
Let's create these data writer functions:
import { CheckpointWriters } from "@snapshot-labs/checkpoint";
import { getAddress } from '@ethersproject/address';
const writers: CheckpointWriters = {
// handleDeploy will get invoked when a contract deployment
// is found at a block
handleDeploy: async (args) => {
// we won't do anything at this time.
// handleNewPost will get invoked when a `new_post` event
// is found at a block
handleNewPost: async ({ mysql, receipt, block }) => {
const event =[0] as any;
// extract posters address from events data
const author = getAddress(BigNumber.from([0]).toHexString());
// post object matches fields of Post entity we will
// define in graphql schema
const posts = {
id: `${author}/${receipt.transaction_hash}`,
created_at_block: block.blockNumber
// table names are `lowercase(EntityName)s` and can be interacted with sql
await mysql.queryAsync('INSERT IGNORE INTO posts SET ?', [post]);
With the above code snippet, we have a data writer that writes new posts to the MySQL database.
You can view a more comprehensive data writer example in our checkpoint-template codebase here.

Starting Checkpoint with Arguments

Finally, we can initialize a checkpoint instance with our arguments like these:
import Checkpoint from "@snapshot-labs/checkpoint";
const checkpoint = new Checkpoint(config, writer, schema);
Next, we start up checkpoint's indexer like this:
The above code will start the checkpoint indexer, and it will begin processing each StarkNet block. When a relevant contract event is found, it gets passed to the data writer for that event.

Querying data

Once Checkpoint has run for a while and has written some data to its database, you can query this data using the generated GraphQL API.
You can mount the query endpoint on any port you like using the graphql handler exported by Checkpoint's object. Like this:
import express from 'express';
const checkpoint = new Checkpoint(...);
const app = express();
app.use('/graphql', checkpoint.graphql);
const PORT = 3000;
app.listen(PORT, () => console.log(`Listening at http://localhost:${PORT}`));
This will mount a GraphQL endpoint that can be accessible at http://localhost:3000/graphql (and a GraphiQL interface at http://localhost:3000/graphql when visited from the browser).
Checkpoint exposes two types of queries:
These enable public users to fetch information about being indexed.
For starters, you can visit http://localhost:3000/graphql in your browser and try running the sample query generated in the graphiql UI.


At this juncture, you should have Checkpoint running, indexing your contracts data and serving this indexed data via graphql.
Next up, you can explore our template repository here to get up and running quickly.
You can also explore some Checkpoint core concepts or move on to the next guide.