Quickstart

This quickstart scaffolds the current EVM template, indexes ETH burn transactions, stores them in PostgreSQL, and exposes them through GraphQL.

Prerequisites

Node.js 20 or newer
Git
Docker with Compose

No data credentials are needed: the template streams data from a public SQD Network Portal dataset (see the EVM networks list).

Install the Squid CLI

npm i -g @subsquid/cli@latest
sqd --version

The version output has the form @subsquid/cli/<version>.

Scaffold the current EVM template

sqd init hello-squid -t evm
cd hello-squid
npm i

The evm alias is maintained by the CLI. It currently retrieves subsquid-labs/squid-evm-template.The project has this layout:

hello-squid

abi

src

abi

main.ts

model

generated

index.ts

.env

.squidignore

commands.json

docker-compose.yml

package.json

schema.graphql

squid.yaml

tsconfig.json

src/main.ts: the data source and the batch handler.
schema.graphql: the data model. sqd codegen generates the TypeORM classes in src/model/generated/ from it.
commands.json: the scripts behind sqd build, sqd up, and sqd down.
squid.yaml: the manifest that sqd run . starts.
db/migrations/: database migrations, applied before the processor starts.
abi/: the place for contract ABI JSON files. sqd typegen generates TypeScript modules from them in src/abi/.

Build the project

sqd build

The template defines build in commands.json. The command generates JavaScript in lib/ and stops on TypeScript errors.

Start PostgreSQL

sqd up

This runs the template’s Docker Compose service with the connection settings from .env.

Run the processor and GraphQL server

sqd run .

The CLI starts the processor and GraphQL services declared in squid.yaml. The processor command applies pending database migrations before it starts.The template’s src/main.ts builds a data source with DataSourceBuilder from @subsquid/evm-stream, requesting transactions sent to the zero address:

src/main.ts

const dataSource = new DataSourceBuilder()
  .setPortal('https://portal.sqd.dev/datasets/ethereum-mainnet')
  .setFields({
    block: {timestamp: true},
    transaction: {hash: true, from: true, value: true},
  })
  .addTransaction({
    where: {to: ['0x0000000000000000000000000000000000000000']},
  })
  .build()

The run() call from @subsquid/batch-processor then drives the batch handler, which turns each transaction into a Burn entity and inserts the batch through ctx.store. The Portal stream covers real-time data as well, all the way to the unfinalized chain head — no additional endpoints are needed.

Query indexed data

Open http://localhost:4350/graphql and run:

query RecentBurns {
  burns(limit: 10, orderBy: block_DESC) {
    id
    block
    address
    value
    txHash
  }
}

Stop the CLI with Ctrl+C. Remove the local database container with:

sqd down

What to change next

Edit the DataSourceBuilder configuration in src/main.ts to select the transactions, logs, traces, or state diffs your application needs. Remember to extend the field selection with every field your handler reads.
Edit schema.graphql, then run sqd codegen and sqd migration:generate after model changes.
Indexing a network without a Portal dataset? Build the source with the EVM RPC source instead. For resilience to endpoint outages, combine Portal and RPC with the EVM fallback source — see High availability.
Migrating an existing pre-Portal squid instead? Follow the Portal migration guide.

Overview

Pipes SDK

Squid SDK

Alternative clients

Prerequisites

What to change next

​Prerequisites

​What to change next

Prerequisites

What to change next