April: The AI That Listens

Aniruddha Ghosh — Sun, 12 Apr 2026 14:14:36 +0000

This is a submission for the DEV April Fools Challenge (https://dev.to/challenges/aprilfools-2026)

What I Built

I built April, a chatbot that doesn’t chat.

April looks like a normal assistant: chat window, typing indicator, blinking eyes. You type. You wait. You try again. April never replies. She only watches and listens.

The joke is simple: we’ve been trained to expect instant answers from AI. April breaks that expectation by doing nothing at all, while still feeling alive through her eyes and subtle animations.

Demo

Demo video:

April’s behavior:

The eyes follow your cursor.
When you type, the eyes track your typing.
When you send a message, April appears to think… forever.
No replies. Ever.

Code

Source code: [https://github.com/AniruddhaGhosh64/april-listens]

How I Built It

I built April using Google Antigravity and vibe-coding my way through the UI.

Frontend: HTML, CSS, JavaScript (generated and iterated using Antigravity)
Animations: CSS + JavaScript for eye tracking, blinking, and typing animations
Logic: Purely client-side behavior. April never sends a response.
No backend and no external hosting. This runs locally.

I focused on interaction design rather than infrastructure. The goal was to make something that feels intelligent without actually being intelligent.

Prize Category

I’m submitting this for the Community Favorite category.

The idea isn’t technical complexity, but emotional reaction. People laugh, then get uncomfortable, then try again anyway. The joke only works when someone interacts with it.

April listens. Always.

Notion Workspace Operating System (Workspace OS)

Aniruddha Ghosh — Sun, 29 Mar 2026 17:37:42 +0000

🚀 Notion MCP Challenge Submission
🧠 What I Built

I built a Notion Workspace Operating System (Workspace OS) powered by Notion MCP + n8n automation, deployed on DigitalOcean.

This system transforms Notion from a passive note-taking tool into an active, intelligent workflow engine that can:

Automatically process incoming data (tasks, notes, ideas)
Categorize and structure information dynamically
Trigger workflows and actions based on context
Maintain a centralized “second brain” with minimal manual effort

🔥 Core Idea

Instead of manually organizing everything in Notion like it’s 2018, this system:

Captures → Processes → Organizes → Acts automatically using MCP + automation pipelines.

🎥 Video Demo

(Insert your Loom / YouTube link here)

Suggested structure for your demo:

Show Notion workspace (Tasks, Projects, Docs DBs)
Add a new entry (manual input/webhook trigger)
Show n8n workflow executing
Show automated categorization + linking in Notion
Highlight MCP interaction (context-aware structuring)

*💻 Show Us the Code
*
🔗 GitHub Repository: (insert your repo link here)

*Key Components:
*

n8n Workflows
Webhook trigger
Data processing nodes (Function / Python)
Notion API nodes
MCP Integration Layer
Context-aware structuring
Smart routing of entries (Tasks / Docs / Projects)
Deployment
Hosted on DigitalOcean (Droplet)
Docker-based setup for n8n

*Tech Stack:
*- Notion API + MCP

n8n (Dockerized)
DigitalOcean (Cloud Hosting)
Python (for logic processing)
GitHub (Version Control)

*⚙️ How I Used Notion MCP
*
Notion MCP is the brain of the system. Instead of treating Notion as static storage, MCP enables:

*🧩 1. Context-Aware Structuring
*
Incoming data is intelligently classified into:

Tasks
Projects
Knowledge Docs

MCP helps interpret intent + context, not just raw input.

*🔄 2. Dynamic Workflow Automation
*
Using MCP + n8n:

Entries trigger workflows automatically
Relationships between databases are created dynamically
Metadata (tags, priority, category) is auto-assigned

*🧠 3. Intelligent Second Brain
*
The system behaves like a lightweight AI assistant:

Organizes information without manual tagging
Maintains clean database relationships
Reduces cognitive load

*⚡ 4. Real-Time Processing Pipeline
*
Flow:

Input (manual/webhook)
n8n triggers workflow
MCP processes context
Data is structured + inserted into Notion
Linked across relevant databases

🏗️ Architecture Overview
Frontend: Notion Workspace (UI layer)
Automation Engine: n8n (self-hosted via Docker)
Intelligence Layer: MCP (context + structuring)
Cloud Hosting: DigitalOcean

💡 Key Features
✅ Automated task & knowledge management
✅ Context-aware classification (via MCP)
✅ Zero manual organization workflow
✅ Scalable cloud-based automation
✅ Clean relational Notion architecture

🚧 Challenges & Solutions

Notion API Structuring Problem: Handling relational fields dynamically Solution: Built flexible mapping logic in n8n
Workflow Failures After Splitting Nodes Problem: Execution errors in multi-step pipelines Solution: Simplified architecture + better data flow handling
Deployment Issues Problem: Initial hosting limitations Solution: Migrated to DigitalOcean using student credits + Docker

**🌱 What’s Next
**Add AI summarization for notes
Implement natural language task creation
Build a dashboard for workflow monitoring
Expand MCP usage for deeper reasoning

*🏁 Final Thoughts
*
This project turns Notion into a true productivity OS instead of just a workspace.

By combining:

Notion MCP (intelligence)
n8n (automation)
Cloud deployment (scalability)

…I created a system that thinks, organizes, and acts with minimal human effort.