Notion Deploy Integration

Deploy Notion-Integrated Applications

Overview

Ship Node.js apps that talk to the Notion API to Vercel, Railway, or Fly.io. This skill covers environment variable management, the Notion client singleton pattern for serverless, rate limit handling at 3 req/sec, health check endpoints that verify Notion connectivity, and caching strategies to reduce API calls.

Deep code lives in references/ so this file stays a lean walkthrough. Read a reference when a step needs the full module, then Write or Edit the code into your project's src/.

Prerequisites

• Node.js >= 18 project with @notionhq/client installed (npm i @notionhq/client)
• Working Notion integration tested locally with a valid NOTION_TOKEN (starts with ntn_)
• Platform CLI installed for your target: vercel, railway, or fly
• Database or page IDs your integration needs access to

Authentication

Every request authenticates with an internal integration token (NOTION_TOKEN, prefix ntn_), created at notion.so/my-integrations and passed as auth to the client. In production the token is stored as an encrypted platform secret and injected at runtime — never committed to source. Store it with vercel env add NOTION_TOKEN production, railway variables set NOTION_TOKEN=..., or fly secrets set NOTION_TOKEN=.... Each database or page must also be explicitly shared with the integration in the Notion UI, or queries return ObjectNotFound.

Instructions

Step 1 — Prepare the application for production

Build a production entry point with four modules: a Notion client singleton, a rate limiter, a response cache, and a health check. The singleton is the essential piece — serverless containers recycle unpredictably, so a module-level client reuses connections across warm invocations instead of paying cold-start and rate-limit cost per request:

// src/notion-client.ts — singleton for serverless environments
import { Client, LogLevel } from '@notionhq/client';

let client: Client | null = null;

export function getNotionClient(): Client {
  if (!client) {
    if (!process.env.NOTION_TOKEN) {
      throw new Error('NOTION_TOKEN environment variable is not set');
    }
    client = new Client({
      auth: process.env.NOTION_TOKEN,
      logLevel: process.env.NODE_ENV === 'production' ? LogLevel.WARN : LogLevel.DEBUG,
      timeoutMs: 30_000,
    });
  }
  return client;
}

The rate limiter (token bucket for the 3 req/sec cap), the TTL response cache, and the healthCheck() function are provided in full — read production-ready application modules and copy each into src/.

Step 2 — Deploy to your target platform

Pick one platform. All three inject NOTION_TOKEN at runtime from an encrypted secret:

• Vercel — serverless functions; best for Next.js apps, API routes, low-traffic webhooks (~200ms cold start).
• Railway — always-on containers; best for long-running sync services and apps needing persistent state.
• Fly.io — edge containers; best for global distribution and multi-region low-latency proxies.

Each path (secret setup, deploy command, and the framework-specific API-route wiring for the singleton, rate limiter, and cache) is in platform deployment paths.

Step 3 — Monitor Notion API errors in production

Add structured error logging so Notion-specific failures surface in your monitoring tool (Sentry, Datadog, or platform logs). The full classifyNotionError / logNotionError module — which maps every @notionhq/client error code to a retryability flag and an operator action, plus the API-route wiring — lives in production error monitoring. Copy it to src/notion-error-handler.ts and call logNotionError(error, context) from every catch block that touches the Notion API.

Key metrics to watch:

• Rate limit hits (429 responses) per minute — alert if sustained above 5/min
• Health check latency — alert if Notion latencyMs exceeds 2000ms
• Auth failures (401/403) — alert immediately; means token rotation is needed
• Cache hit ratio — target > 70% in steady state; a low ratio means wasted API calls

Output

This workflow produces:

• Node.js application deployed to Vercel, Railway, or Fly.io
• NOTION_TOKEN stored as an encrypted platform secret (never in source code)
• Notion client singleton that reuses connections across serverless invocations
• Rate limiter enforcing the 3 req/sec Notion API limit
• In-memory response cache reducing redundant API calls
• /health endpoint that verifies live Notion API connectivity
• Structured error logging classifying Notion API failures by severity

Error Handling

Examples

Two complete, copy-ready examples live in full deployment examples:

• Minimal Express server — a platform-agnostic server wiring the singleton, rate limiter, health check, and error handler into /health and /api/query routes.
• Deploy script — a deploy.sh [vercel|railway|fly] that builds, sets the secret, deploys, and verifies the health endpoint.

Resources

• Notion API Reference — official REST API docs
• @notionhq/client npm — official SDK with built-in retry
• Notion API Rate Limits — 3 req/sec per integration
• Notion API Status — check during outages
• Vercel Environment Variables — secret management
• Railway Variables — encrypted secrets
• Fly.io Secrets — runtime secret injection

Next Steps

• Add webhook receivers with the notion-webhooks-events skill
• Set up database sync pipelines with the notion-sync-databases skill
• Implement page content extraction with the notion-extract-content skill

Source: jeremylongshore/claude-code-plugins-plus-skills → skills/.curated/notion-deploy-integration/SKILL.md

Also appears in: jeremylongshore/claude-code-plugins-plus-skills/plugins/saas-packs/notion-pack/skills/notion-deploy-integration/SKILL.md