🤖 Daily GenAI & Agentic AI Career Briefing

An automated n8n pipeline that wakes up every morning, scrapes the latest GenAI & Agentic AI news from two live sources, deduplicates them using stateful memory, lets Google Gemini extract a structured JSON briefing, and delivers a beautiful HTML email digest directly to your Gmail inbox — zero clicks required.

---

⚡ How It Works

↑ End-to-end 6-step automation pipeline — pencil sketch overview

The workflow runs automatically every day via a cron schedule. Here's what happens under the hood:

Step	Node	What It Does
1	Daily Trigger (Cron)	Fires at 8:00 AM daily. Two parallel triggers kick off the Google News and Hacker News branches simultaneously.
2	Fetch Google News RSS	Hits Google News RSS with a targeted query: "agentic AI" OR "generative AI" OR "AI agent" OR "AI tool" OR "AI regulation"
2	Fetch Hacker News	Calls the Hacker News Algolia API (search_by_date?query=AI agent) to pull the 20 most recent stories.
3	Combine → Dedupe & Filter	Merges both feeds. A custom JavaScript node checks workflow static data to skip already-seen URLs, filters to the last 24 hours, sorts by recency, and keeps the top 15 freshest articles.
4	Has New Items?	A gatekeeping filter node. If count == 0, the workflow stops immediately — no blank emails sent.
5	Build Career Briefing (Gemini)	Feeds the 15 headlines into gemini-3.1-flash-lite with a strict structured JSON prompt. Extracts: Market Sentiment, Top News + Impact, Models & Tools, Skills to Learn, Industry Impact.
6	Format HTML Email	A JavaScript Code Node parses the JSON and compiles a clean, fully-styled HTML email template.
7	Send Email Digest	Dispatches the HTML email via Gmail OAuth SMTP to your inbox.

---

🧠 AI Engineering Highlights

This project demonstrates production-grade AI engineering patterns beyond simple prompting:

✅ Structured LLM Outputs (JSON Schema Enforcement)

Instead of relying on fragile free-text generation, Gemini is constrained via a strict JSON schema prompt. The output schema is enforced at the prompt level:

{
  "sentiment":       { "score": "Bullish | Bearish | Neutral", "reason": "..." },
  "top_news":        [{ "headline": "...", "impact": "..." }],
  "models_tools":    ["tool1", "tool2"],
  "skills":          [{ "skill": "...", "reason": "..." }],
  "industry_impact": "..."
}

A downstream JavaScript node strips any leaked markdown fences and JSON.parse()s safely.

✅ Stateful Deduplication via Workflow Memory

The deduplication node uses $getWorkflowStaticData('global') — n8n's built-in stateful storage — to maintain a rolling list of up to 500 sent article URLs across runs. This means you never receive the same article twice, even across multiple days.

✅ Separation of Concerns

AI processing is strictly separated from presentation logic:

• 🤖 LLM → handles only data extraction
• 🖥️ JS Code Node → handles all HTML templating

✅ Gatekeeping Filter (No Blank Emails)

Before invoking Gemini, the workflow checks count > 0. If there are no fresh articles (e.g., a weekend with low news volume), execution stops and no API calls are made.

✅ Market Sentiment Analysis

The AI is used as an analytical engine to grade the overall tone of today's AI news as Bullish, Bearish, or Neutral, with a one-sentence rationale.

---

🏗️ System Architecture

↑ Node routing, parallel fetch branches, decision gatekeeping, and AI synthesis — pencil sketch

Node Map

[Daily Trigger - News] ──→ [Fetch Google News RSS] ──┐
                                                       ├──→ [Combine News Sources]
[Daily Trigger - HN]   ──→ [Fetch Hacker News]  ──→  │
                              [Split HN Hits]   ──────┘
                                                       ↓
                                          [Dedupe, Filter & Build Digest]
                                                       ↓
                                            [Has New Items? (Filter)]
                                                       ↓ YES
                                          [Build Career Briefing]  ←── [Gemini Model]
                                                       ↓
                                          [Format HTML Email (JS)]
                                                       ↓
                                           [Send Email Digest (Gmail)]

---

📧 What the Email Looks Like

The delivered HTML email contains 5 structured sections:

Section	Content
Market Sentiment	Bullish / Bearish / Neutral + one-line rationale
🔥 Top News Today	Up to 5 headlines with impact summary per article
🤖 New Models & Tools	Extracted tool names mentioned in today's news
🛠️ Skills to Improve	Recommended skills with context on why they're trending
💼 Industry Impact	2-3 sentence analysis of what this means for engineering roles

---

⚡ Quick Start & Import

↑ 4-step setup — up and running in under 5 minutes

Step 1: Get n8n Running

Option A — Docker (Recommended)

# Copy environment template
cp .env.example .env
# Fill in your values in .env, then:
docker compose up -d
# Open: http://localhost:5678

Option B — n8n Cloud Sign up at n8n.io and use the cloud editor directly.

---

Step 2: Import the Workflow

1. Download genai-daily-career-briefing.json
2. In n8n, click + (top-left) → Import from File
3. Upload the .json file

---

Step 3: Set Credentials

You need to configure 2 credentials inside n8n:

Gemini API Key

• Go to Google AI Studio → Create API Key
• In n8n: open the Gemini Model node → paste your key

Gmail OAuth

• Set up a Google Cloud project and OAuth 2.0 credentials
• Guide: n8n Google OAuth docs
• In n8n: open Send Email Digest node → connect your Gmail account

---

Step 4: Set Your Recipient Email

• Open the Send Email Digest node
• Change YOUR_EMAIL@gmail.com in the sendTo field to your actual email

---

Step 5: Activate!

• Toggle the workflow to Active (top-right switch)
• (Optional) Click Execute Workflow to trigger an immediate test run

---

🔐 Configuration & Security

# 1. Copy the template
cp .env.example .env

# 2. Fill in your values
GEMINI_API_KEY=your_gemini_api_key_here
RECIPIENT_EMAIL=your_email@gmail.com
GMAIL_CLIENT_ID=your_google_client_id_here
GMAIL_CLIENT_SECRET=your_google_client_secret_here

⚠️ The .env file is excluded from Git via .gitignore. Never commit real credentials.

---

📁 Repository Structure

genai-n8n/
├── genai-daily-career-briefing.json   # ← Complete n8n workflow (import this)
├── docker-compose.yml                 # Self-hosted n8n via Docker
├── .env.example                       # Environment variable template
├── .gitignore                         # Ignores .env and sensitive files
├── README.md                          # This file
└── assets/
    ├── how_it_works_sketch.png        # Pipeline overview (pencil sketch)
    ├── architecture_sketch.png        # System architecture (pencil sketch)
    ├── setup_steps.png                # Quick start visual (pencil sketch)
    ├── future_roadmap.png             # Roadmap diagram (pencil sketch)
    ├── email_preview.png              # Sample email output screenshot
    └── hero_banner.png                # Project banner image

---

🏷️ Project Tags

Area	Tags
Automation
Artificial Intelligence
Integrations
Career & Learning
Dev Style

---

🚀 Future Improvements

↑ Planned evolution from single-agent pipeline to full multi-agent system — pencil sketch

Phase 1 — Core Extensions

Improvement	Description
🗄️ Vector Database (RAG)	Connect to Pinecone or Qdrant to store historical article embeddings. The LLM can then reference past context to detect long-term industry trends and avoid repeating insights from previous briefings.
🕷️ Full Article Scraping	Replace RSS summary text with a Puppeteer / HTTP Request node that fetches the full article body before passing it to Gemini — enabling much deeper, nuanced analysis instead of headline-only summaries.
📰 More News Sources	Add Dev.to, Towards Data Science, ArXiv AI abstracts, or LinkedIn Pulse as additional RSS/API sources to broaden coverage.

Phase 2 — Multi-Agent Architecture

Improvement	Description
🔍 Researcher Agent	Use n8n's LangChain agent nodes to deploy a dedicated Researcher Agent that autonomously decides which sources to query, ranks articles by relevance, and summarizes findings for the Writer.
✍️ Writer Agent	A separate Writer Agent receives the ranked research and drafts a personalized briefing with a consistent editorial voice — using memory of your past preferences.
✅ Editor / Fact-Check Agent	A final verification agent that cross-references claims in the briefing against a trusted knowledge base before the email is dispatched.

Phase 3 — Multi-Channel Delivery

Improvement	Description
💬 Slack / Discord Bot	Add a Webhook output node to push the briefing as a formatted message to your team's Slack or Discord channel every morning.
🎙️ Audio Podcast Digest	Pipe the briefing text through a Text-to-Speech node (ElevenLabs or Google TTS) to generate a short audio podcast you can listen to on your commute.
📊 Web Dashboard	Build a lightweight Next.js frontend that reads briefing history from a database and visualizes sentiment trends, trending tools, and skill frequency over time.
🌍 Multi-Language Support	Add a translation step after the AI parser to deliver the briefing in the user's preferred language using Gemini's translation capabilities.
📅 Weekly Summary	Add a separate Saturday workflow that aggregates the week's 5 briefings into a single executive summary email with trend highlights.

---

🤝 Contributing

1. Fork the repository
2. Create a feature branch: git checkout -b feature/my-improvement
3. Commit your changes: git commit -m "feat: add Slack delivery channel"
4. Push and open a Pull Request

Feedback and workflow forks are welcome! If you add a new source or delivery channel, please share the .json workflow export.

---

I WILL BUILT WITH PASSION WITH INTREST IN AI.