Connect your SDKs

By default every Pug SDK sends data to Pug Cloud (https://api.pugs.dev). When you self-host, you change one thing: point the SDK’s endpoint at your own pug server. Nothing else about the API changes - same methods, same events. Keys don’t carry over from Pug Cloud: your instance issues its own, so create a project in your dashboard and use its project ID and keys.

Use the public URL of your server (the TLS-terminated origin in front of port 3000, e.g. https://api.example.com), not the internal :3000 address.

The one setting, per SDK

It’s the same option, endpoint, on every SDK (the HTTP API just uses the request URL):

Browser

import { init } from '@pug-sh/browser'

init('YOUR_PROJECT_ID', {
  apiKey: 'pub_YOUR_PUBLIC_KEY',
  endpoint: 'https://api.example.com', // your pug server
})

For environment-aware config (e.g. Vite), fall back to localhost in dev:

init(import.meta.env.VITE_PUG_PROJECT_ID, {
  apiKey: import.meta.env.VITE_PUG_PUBLIC_KEY,
  endpoint: import.meta.env.DEV ? 'http://localhost:3000' : 'https://api.example.com',
})

The one-tag CDN install takes it as data-endpoint="https://api.example.com" - see One-tag install.

Flutter

import 'package:pug_flutter/pug_flutter.dart';

await Pug.init(
  'YOUR_PROJECT_ID',
  const PugOptions(
    apiKey: 'pub_your_public_key',
    endpoint: 'https://api.example.com', // your pug server
  ),
);

Node (server-side)

The Node SDK takes the same endpoint option and authenticates with your private key (prv_...):

import { Pug } from '@pug-sh/node'

const pug = new Pug({
  apiKey: process.env.PUG_PRIVATE_KEY!, // prv_... - server-side only
  endpoint: 'https://api.example.com', // your pug server
})

Ruby (server-side)

The Ruby SDK takes the same endpoint option and authenticates with your private key (prv_...):

require "pug"

pug = Pug::Client.new(
  api_key: ENV.fetch("PUG_PRIVATE_KEY"), # prv_... - server-side only
  endpoint: "https://api.example.com",   # your pug server
)

It doesn’t follow redirects, so use the URL your server answers on: an http:// origin that redirects to https:// fails every request as unimplemented, and dropped events go to on_error.

HTTP API

If you call the API directly, just swap the host in the request URL - the paths are unchanged. The key identifies the project, so the body carries only the events (HTTP API):

curl -X POST https://api.example.com/sdk.events.v1.EventsService/BatchCreate \
  -H 'Content-Type: application/json' \
  -H 'Connect-Protocol-Version: 1' \
  -H 'x-api-key: pub_YOUR_PUBLIC_KEY' \
  -d '{ "events": [ /* ... */ ] }'

Notes

• No SDK-specific CORS setup is needed on the server. The events and profiles ingest services accept any origin (auth lives entirely in the x-api-key header). Only the dashboard origin needs listing in PUG_CORS_ORIGINS - see Dashboard.
• Key types are unchanged. Public keys (pub_...) for the browser/Flutter SDKs, private keys (prv_...) for the Node and Ruby SDKs and server-to-server reads. See Authentication.
• Verify it works: open your self-hosted dashboard’s Live view - an event fired from a repointed SDK should appear within ~15 seconds. If requests fail with a network error, confirm the endpoint is your public HTTPS origin and that TLS terminates in front of the server.

For the full initialization options (batching, sessions, auto-capture, consent), see the SDKs reference.

Further reading

• Dashboard: the dashboard’s own VITE_API_BASE_URL endpoint setting
• Deployment: stand up the server your SDKs point at
• Development: point SDKs at a local pug dev