| [ Web Proxy ] |
| Viewing: https://quicknode.com/docs/quicknode-sdk | [Back] [Original] |
The SDK is a single SDK for working with Quicknode product APIs from the language you already use: Rust, Python, Node.js, and Ruby.
Use it to build Quicknode product workflows, operational services, scripts, and AI agents that need typed access to:
Complete reference for developers and agents building with the SDK:
Use the SDK when you want one consistent client for Quicknode product APIs, especially when your app or agent needs to coordinate multiple products in one workflow.
Common workflows include:
| Language | Package | Install |
|---|---|---|
| Node.js / TypeScript | @quicknode/sdk | npm install @quicknode/sdk |
| Python | quicknode-sdk | pip install quicknode-sdk |
| Rust | quicknode-sdk | cargo add quicknode-sdk --features rust |
| Ruby | quicknode_sdk | gem install quicknode_sdk |
The SDK is built around a shared Rust core with bindings for specific languages. The core compiles to native libraries for each supported platform, giving you predictable performance and a small dependency footprint. The trade-off is that we publish binaries for a specific set of targets rather than running in every environment our host languages support.
Precompiled native modules are published for:
| Platform | Targets |
|---|---|
| Linux (glibc) | x86_64, aarch64 on glibc 2.17+ (manylinux2014) |
| Linux (musl) | x86_64, aarch64 on Alpine and other musl distros |
| macOS | Apple Silicon (arm64) |
Linux glibc binaries are built against glibc 2.17, so they load on any distro released from 2014 onward, including RHEL 7+, Ubuntu 14.04+, Debian 8+, Amazon Linux 2+, SLES 12+, and Fedora 19+.
On an unsupported platform, importing the SDK fails fast at load time with an error listing the available targets. The failure surfaces at install or import, not at first call.
Construct the SDK once, then use the product clients exposed from that shared configuration.
| Client | Purpose | REST API |
|---|---|---|
admin | Endpoints, teams, usage, logs, billing, metrics, security, and rate limits | https://api.quicknode.com/v0/ |
rpc | On-chain JSON-RPC calls via Tooling Access or x402/MPP micropayments | Tooling Access, x402, or MPP gateway |
streams | Streams creation, updates, lifecycle, and filter testing | https://api.quicknode.com/streams/rest/v1/ |
webhooks | Webhook templates, destinations, lifecycle, and counts | https://api.quicknode.com/webhooks/rest/v1/ |
kvstore | Sets and lists for persisted state | https://api.quicknode.com/kv/rest/v1/ |
sql | SQL queries against indexed blockchain data | https://api.quicknode.com/sql/rest/v1/ |
The SDK entry point is QuicknodeSdk in each language binding. Ruby exposes it as QuicknodeSdk::SDK.
The sql client provides two methods:
query(sql, clusterId): Execute a SQL query against indexed blockchain data. Returns typed rows and query statistics (credits, rows scanned, bytes read).getSchema(clusterId): Retrieve the table schema for a cluster.const qn = new QuicknodeSdk();
// Run a SQL query
const resp = await qn.sql.query(
"SELECT action_type, user FROM hyperliquid_system_actions LIMIT 3",
"hyperliquid-core-mainnet"
);
console.log(`${resp.rows} rows, ${resp.credits} credits`);
// Get the schema for a cluster
const schema = await qn.sql.getSchema("hyperliquid-core-mainnet");
Create an API key in the Quicknode dashboard, then configure the SDK with the QN_SDK__API_KEY environment variable.
export QN_SDK__API_KEY="YOUR_API_KEY"
The SDK also supports base URL overrides for local development, staging, and agent sandboxes:
| Environment variable | Default |
|---|---|
QN_SDK__ADMIN__BASE_URL | https://api.quicknode.com/v0/ |
QN_SDK__STREAMS__BASE_URL | https://api.quicknode.com/streams/rest/v1/ |
QN_SDK__WEBHOOKS__BASE_URL | https://api.quicknode.com/webhooks/rest/v1/ |
QN_SDK__KVSTORE__BASE_URL | https://api.quicknode.com/kv/rest/v1/ |
QN_SDK__SQL__BASE_URL | https://api.quicknode.com/sql/rest/v1/ |
QN_SDK__HTTP__TIMEOUT_SECS | 30 |
The SDK is designed to be useful for humans and agents. Agents can initialize one SDK handle, reuse the same credentials across product clients, and rely on typed method inputs and responses instead of composing raw REST calls for each workflow.
| Web Proxy Viewer | New URL | Original Page |