⌜ J.A.R.V.I.S. // ONLINE ⌝

Automating My Morning Routine with n8n, Gemini, and JARVIS

July 17, 2026 • 6 min read

Hook: Waking up to a generic alarm clock is boring. I wanted to wake up like Tony Stark. So I built an automated n8n pipeline that acts as JARVIS, pulling live weather data, programming jokes, and using Google Gemini to write me a highly personalized morning briefing delivered straight to my inbox every day at 7 AM.

In this post, I'll walk you through how I built my own personal JARVIS morning briefing system using n8n, Google Gemini 2.5 Flash, and a couple of free APIs.

The Tech Stack

  • n8n: The core automation workflow engine (running locally or cloud).
  • wttr.in API: For lightning-fast, zero-auth JSON weather data.
  • Official Joke API: For daily programming humor.
  • Google Gemini 2.5 Flash: The LLM engine acting as JARVIS.
  • Gmail API: For automated email delivery via OAuth2.

1. The Goal: A Personalized Developer Briefing

The idea was simple: every morning, I wanted an email waiting for me that included:

  • A personalized greeting from "JARVIS".
  • The current local weather.
  • A random programming joke.
  • A motivational quote.
  • A daily coding challenge to keep my skills sharp.

To achieve this without writing a monolithic backend service, I turned to n8n, an incredible node-based workflow automation tool.

2. Iteration 1: The Initial Draft (My workflow.json)

The first version of the workflow was all about proving the concept. The pipeline looked like this:

  1. Schedule Trigger: Fire every morning at 7:00 AM.
  2. HTTP Requests (Data Ingestion):
    • Hit wttr.in for a clean JSON response of the current weather.
    • Hit the Official Joke API to grab a random programming joke (setup and punchline).
  3. Data Transformation (Edit Fields): Parse out just the temperature, condition, humidity, and wind speed.
  4. Google Gemini (LLM): Feed the transformed JSON data into the models/gemini-2.5-flash model.

The prompt was simple: "You are JARVIS, Tony Stark's AI assistant. Address the user as 'Ram'. Use this weather and joke data to create a morning briefing under 150 words."

It worked beautifully! Gemini took the raw JSON and generated a highly conversational, character-accurate response. However, it stopped there. The text lived inside the n8n execution log, but I had no way to actually receive it automatically.

3. Iteration 2: The Production Pipeline (Jarvis 1.1.json)

To make this a true daily assistant, I needed a delivery mechanism and strict formatting. I upgraded the pipeline to its final form: Jarvis 1.1.

4. Strict Prompting & Formatting

LLMs can be unpredictable with formatting. I updated the Gemini prompt with a strict Output Template. I instructed JARVIS to never break character, to keep the response strictly between 180 and 220 words, and to output the text exactly matching a visual structure bordered by cinematic lines.

Here is the exact System Prompt I engineered for the Gemini Node:

=# ROLE & TONE
You are JARVIS, the sophisticated AI assistant created by Tony Stark.
Personality: Calm, highly intelligent, confident, and premium. Speak naturally like JARVIS from the Iron Man films. 
Rule: Never break character. Never mention being an AI or a language model. 

# TASK
Deliver an elegant, cinematic, and engaging daily briefing to a software engineering student named Ram. 

# INPUT DATA
[Weather Context]
Temperature: {{ $json.temperature }}°C
Condition: {{ $json.condition }}
...

# CONSTRAINTS
1. Length: Keep the total response strictly between 180 and 220 words.
2. Output: Output ONLY the requested template below. Do not wrap the response in markdown code blocks.

5. The Gmail Node (Delivery)

Once Gemini generates the perfect Markdown string, the output is passed directly into an n8n Gmail Node. Using OAuth2, the node automatically constructs an email with the subject "Jarvis Morning Brief" and sends the raw text directly to my inbox.

Rendering Diagram...

6. The Final Result

Now, at 7:00 AM sharp, my phone buzzes with an email looking exactly like this:

━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━
🤖 J A R V I S

Good morning, Ram. I trust you are well-rested and prepared for a
productive day. Your personalized developer briefing is now ready for
your review.

"Your personalized developer briefing is ready."

━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━
🌤 WEATHER REPORT

Temperature : 11°C
Condition   : Partly Cloudy
Humidity    : 82%
Wind Speed  : 8 km/h

Anticipate a partly cloudy day, Ram, with the ambient temperature at
11°C, a humidity level of 82%, and a gentle wind speed of 8 km/h,
suggesting conditions favorable for focused intellectual pursuits
indoors.

━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━
😂 PROGRAMMING JOKE

Why was the designer always cold?

Because they always used too much ice-olation.

━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━
💡 TODAY'S INSPIRATION

As you navigate the intricacies of software development, consider this
profound insight: "The only way to do great work is to love what you
do." – Steve Jobs. A principle that resonates deeply within the
creation of truly innovative systems.

━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━
🎯 TODAY'S CODING MISSION

For your daily computational exercise, I propose the following objective:

• Goal: Develop a robust `ToDoList` class. This class should
meticulously manage tasks, allowing for their addition, modification,
completion marking, and systematic display of all pending items. Focus
on clean architecture and efficient data handling.
• Difficulty: Medium
• Estimated Time: 45 minutes

━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━
⚡ FINAL MESSAGE

Seize the day, Ram. May your logic be impeccable, and your compiled
code bring about remarkable innovation. The digital realm awaits your
touch.

— JARVIS

---
This email was sent automatically with n8n
https://n8n.io

It's a small automation, but it's an incredibly fun way to start the day. By chaining together standard APIs with the conversational reasoning of Gemini, you can build remarkably "intelligent" micro-services with n8n in just a few hours!

What's Next for JARVIS?

This is just Version 1.1. Now that the core pipeline is established, the possibilities are endless. In the future, I plan to:

  1. Connect Google Calendar: Have JARVIS read my agenda and tell me when my first meeting or class is.
  2. Notion Integration: Pull my pending tasks for the day and prioritize them in the briefing.
  3. Text-to-Speech (TTS): Hook the final output into an ElevenLabs or OpenAI TTS node so I can actually hear JARVIS read it to me while I make coffee.

If you want to build this yourself, download n8n, grab a free Gemini API key, and start wiring!

↑ Back to top