The Daily Read

A stateless, serverless technology news compilation engine that aggregates morning headlines daily, digests summaries using Google Gemini, and emails the digest to Anand's inbox using Resend.

📁 Submodule Project Structure

Located at apps/newsletter-generator:

apps/newsletter-generator/
├── app/
│   ├── api/
│   │   └── cron/
│   │       └── route.ts   # Main scheduled cron route handler
│   ├── layout.tsx
│   └── page.tsx           # Subscriber opt-in visualizer page
├── lib/
│   ├── scraper.ts         # Stateless DuckDuckGo parser wrapper
│   └── templates.ts       # HTML news email template builder
├── package.json
└── tsconfig.json

⚙️ Cron Execution & Security

Bearer Token Security Gate

To prevent unauthorized users from triggering heavy web crawls and invoking AI models, the serverless route is gated with cryptographically strong bearer tokens. The Next.js Edge route intercepts all incoming requests:

Endpoint: POST /api/cron

Request Headers:

Authorization: Bearer <CRON_SECRET>
Content-Type: application/json

Validation Logic:

import { NextRequest, NextResponse } from "next/server";

export async function POST(request: NextRequest) {
  const authHeader = request.headers.get("Authorization");
  if (!authHeader || authHeader !== `Bearer ${process.env.CRON_SECRET}`) {
    return NextResponse.json(
      { error: "Unauthorized access blocked." }, 
      { status: 401 }
    );
  }
  // Proceed with news crawl...
}

🌐 Scraper Proxy Mechanism

Since hosting servers do not maintain a permanent browser instance, the news crawler in lib/scraper.ts executes a lightweight scrape via client-side search proxies: 1. Formulates search query strings (e.g. https://html.duckduckgo.com/html/?q=site:techcrunch.com+OR+site:wired.com+"AI"&s=d). 2. Performs an HTTP GET request with a customized user-agent to bypass basic scrape protection. 3. Parses the HTML response, extracting news headers, links, and publication tags using regular expression selectors, avoiding heavy DOM parsing engines.

🧠 Gemini Fallback Cascade

The generator utilizes the official @google/genai SDK. To guarantee morning news delivery on developer free tiers, the pipeline executes a cascading retry loop if quota boundaries are breached:

import { GoogleGenAI } from "@google/genai";

async function compileDigest(contentPayload: string): Promise<string> {
  const ai = new GoogleGenAI({ apiKey: process.env.GEMINI_API_KEY });
  const modelList = ['gemini-2.5-flash', 'gemini-2.5-flash-lite'];
  let lastError = "";

  for (const model of modelList) {
    try {
      console.log(`Compiling news via: ${model}`);
      const response = await ai.models.generateContent({
        model,
        contents: `Compile a morning tech newsletter. Context:\n${contentPayload}`,
        config: {
          systemInstruction: "You are a professional tech newsletter editor. Output styled Markdown."
        }
      });
      if (response.text) return response.text;
    } catch (e: any) {
      lastError = e.message;
      console.warn(`Model ${model} failed: ${e.message}. Retrying fallback...`);
    }
  }
  throw new Error(`All models failed. Last error: ${lastError}`);
}

✉️ Resend Mailer Payload

Once the news digest is compiled, it is styled into a dark-themed HTML layout using the templates in lib/templates.ts and posted to the Resend API:

Endpoint: https://api.resend.com/emails

JSON Body Parameters:

{
  "from": "Anand's News <newsletter@anandmuraleedharan.com>",
  "to": ["anand@example.com"],
  "subject": "The Daily Read - Morning Tech Digest",
  "html": "<html>...styled email template...</html>"
}

🚀 Local Development Commands

Install Submodule Dependencies:

cd apps/newsletter-generator && npm install

Start Dev Server Locally:
```
npm run dev
```
(Running on http://localhost:3001)

Manually Trigger Cron Execution:

curl -X POST http://localhost:3001/api/cron \
  -H "Authorization: Bearer your_local_secret"