Skip to main content
How to use:

login

Opens your browser to sign in to Goldsky (Google, GitHub, SSO, or email) and authenticates the CLI automatically — there’s no API key to copy. For scripted or headless environments, pass a pre-created API key with --token, or use --no-browser to print the login URL instead of opening a browser. How to use:

logout

How to use:

subgraph

How to use:

subgraph deploy

How to use:

subgraph list

How to use:

subgraph delete

How to use:

subgraph tag

How to use:

subgraph tag create

How to use:

subgraph tag delete

How to use:

subgraph webhook

How to use:

subgraph webhook create

How to use:

subgraph webhook delete

How to use:

subgraph webhook list

How to use:

subgraph webhook list-entities

How to use:

subgraph log

How to use:

subgraph pause

How to use:

subgraph start

How to use:

subgraph update

How to use:

subgraph init

How to use:

project

How to use:

project users

How to use:

project users list

How to use:

project users invite

How to use:

project users remove

How to use:

project users update

How to use:

project leave

How to use:

project list

How to use:

project update

How to use:

project create

How to use:

pipeline

How to use:

pipeline get

How to use:

pipeline export

How to use:

pipeline apply

How to use:

pipeline get-definition

How to use:

pipeline create

How to use:

pipeline update

How to use:

pipeline delete

How to use:

pipeline list

How to use:

pipeline monitor

How to use:

pipeline pause

How to use:

pipeline start

How to use:

pipeline stop

How to use:

pipeline info

How to use:

pipeline resize

How to use:

pipeline validate

How to use:

pipeline cancel-update

How to use:

pipeline restart

How to use:

pipeline snapshots

How to use:

pipeline snapshots list

How to use:

pipeline snapshots create

How to use:

dataset

How to use:

dataset get

How to use:

dataset list

How to use:

indexed

How to use:

indexed sync

How to use:

indexed sync decoded-logs

How to use:

indexed sync raw-blocks

How to use:

indexed sync raw-logs

How to use:

indexed sync raw-transactions

How to use:

secret

How to use:

secret create

How to use:
Running goldsky secret create with no --value launches an interactive prompt that first asks you to select a secret type (jdbc, clickHouse, elasticSearch, opensearch, kafka, s3, sqs, pubsub, dynamodb, or httpauth) and then collects the fields that type requires. Passing both --name and --value (JSON) creates the secret non-interactively and skips the type selector. See Mirror secrets for details on each type and the JSON schema for --value.

secret list

How to use:

secret reveal

How to use:

secret update

How to use:

secret delete

How to use:

hosted-sink

Commands related to Goldsky-hosted databases. Provisions a fully managed Postgres (Neon) database and stores its credentials as a Goldsky secret, so you can use it as a pipeline sink without bringing your own database.
goldsky hosted-sink requires the Scale plan or above. On an account without access, the command fails with a message pointing to your team’s billing page (https://app.goldsky.com/teams/<teamId>/billing).
How to use:

hosted-sink create

Provisions a Goldsky-hosted Postgres (Neon) database and stores its credentials as a secret. When --name is omitted, a name is generated automatically as HOSTED_POSTGRES_<RANDOM>. How to use:
On success the command prints the created secret’s name, ID, and type:
The database connection string is intentionally not printed, so credentials are never written to logs. Reference the database from a pipeline by using the printed secret name as the sink’s secret_name.

edge

Commands related to Edge endpoint management. Create and manage Edge endpoints (API keys) from the terminal, and inspect the networks and datasets Edge exposes. How to use:
Each endpoint belongs to one product: rpc (Edge RPC), data (Edge Data), or boost (Boost). Boost endpoints also have their own command group, goldsky boost, for configuring providers.

edge create

Creates an Edge endpoint. The name is a label you choose; the API key is generated for you and can be printed at any time with edge reveal. How to use:
--rate-limit takes a named budget. Budgets are named edge-tier-<total>-total-<per-ip>-per-ip, where the total is the requests-per-minute cap across all callers of the endpoint and the per-IP part is the cap for any single client IP. Pass an invalid value to print the current list:
When --rate-limit is omitted the endpoint uses your plan’s default budget. --allowed-domains restricts browser callers to the listed web origins (for example https://app.example.com,https://staging.example.com); server-side callers are unaffected. Endpoints are unrestricted by default.

edge list

How to use:
Prints a table with each endpoint’s name, product, status (for example ACTIVE, or PAUSED after edge pause), rate-limit budget, allowed domains, and creation time.

edge get

How to use:
Example output:
The API key is not included; use edge reveal for that.

edge update

How to use:
Pass --rate-limit default to return to your plan’s default budget, and --allowed-domains "" to remove the browser-origin restriction. See edge create for the list of budgets.

edge delete

How to use:
Deleting an endpoint revokes its API key immediately. Any client still using the key gets authentication errors. Use edge pause if you want to stop traffic without losing the key.

edge pause

Stops the endpoint from serving requests while keeping its key and configuration. Requests to a paused endpoint are rejected until it is resumed. How to use:

edge resume

How to use:

edge reveal

Prints the endpoint’s API key. Treat the output as a secret; it is the same key shown in the dashboard. How to use:

edge metrics

Prints request and error counts for the endpoint as JSON, bucketed over the window you choose. Defaults to the last 24 hours. The output has two arrays, requests and errors, one entry per bucket. How to use:

edge networks

Lists the chains Edge RPC serves: the network slug, display name, parent chain, and chain ID. The chain ID is what goes in the endpoint URL (https://edge.goldsky.com/standard/evm/{chainId}?key={your-key}). See supported networks for the same list with per-chain details. How to use:

edge sources

Lists the datasets available through Edge Data (for example blocks/{chain}, blocks/{chain}/head, and the Polymarket datasets), with the provider and a one-line description of each. How to use:

boost

How to use:
Boost is project-scoped, so its commands do not take an endpoint name. provider connect supports QuickNode, Alchemy, Chainstack, and dRPC. The first three create dedicated resources in your provider account; dRPC reuses an existing unrestricted key and creates nothing.

boost enable

How to use:

boost disable

How to use:

boost get

How to use:

boost update

How to use:

boost reveal

How to use:

boost metrics

How to use:

boost requests

How to use:

boost chains

How to use:

boost validate

How to use:

boost set

How to use:

boost provider

How to use:

boost provider connect

How to use:

boost provider disconnect

How to use:

boost provider add

How to use:

boost provider replace

How to use:

boost provider activate

How to use:

boost provider remove

How to use:

boost provider set

How to use:

telemetry

How to use:

telemetry status

How to use:

telemetry enable

How to use:

telemetry disable

How to use:

turbo pipelines

How to use:

turbo apply

How to use:

turbo delete

How to use:

turbo pause

How to use:

turbo list

How to use:

turbo logs

How to use:

turbo validate

How to use:

turbo inspect

How to use:
Use --print / -p to bypass the interactive TUI and stream records directly to stdout. This makes output fully selectable and allows piping to tools like jq:
See Live Inspect - Print mode for more details.

turbo get

How to use:

turbo state

How to use: