DEV Community: Zain K.

Notion MCP Challenge Judges Dashboard

Zain K. — Mon, 30 Mar 2026 05:20:00 +0000

What I Built

I built the Notion Judge's Glassboard — a premium, cinematic "Evaluation OS" explicitly designed to streamline and elevate the judging experience for the Notion MCP Challenge.

With over 180+ incredible submissions, evaluating every entry across multiple criteria (Originality, Complexity, Practicality) can become overwhelming. The Glassboard ingests live submission data (via DEV.to API & Notion MCP) and transforms it into a highly aesthetic, two-stage React SPA:

The Cinematic Hero: A sleek, fully animated dark-mode landing page complete with floating ambient particles and interactive SVG badges that instantly make the project feel premium.
The Pastel Gallery: A smooth, glassmorphism-inspired category dashboard that sorts submissions dynamically, surfacing read times, reaction counts, and direct links natively mapped to the actual DEV.to articles.

This isn't just a basic tracker; it's a bespoke evaluation hub built to make the final judging process as impressive as the projects being evaluated.

DEMO

Show us the code

The entire React/Vite front-end and Python data-ingestion pipeline is available here:

zainkhan1994 / notion-judge-glassboard

React + TypeScript + Vite

This template provides a minimal setup to get React working in Vite with HMR and some ESLint rules.

Currently, two official plugins are available:

@vitejs/plugin-react uses Oxc
@vitejs/plugin-react-swc uses SWC

React Compiler

The React Compiler is not enabled on this template because of its impact on dev & build performances. To add it, see this documentation.

Expanding the ESLint configuration

If you are developing a production application, we recommend updating the configuration to enable type-aware lint rules:

export default defineConfig([
  globalIgnores(['dist']),
  {
    files: ['**/*.{ts,tsx}'],
    extends: [
      // Other configs...
      // Remove tseslint.configs.recommended and replace with this
      tseslint.configs.recommendedTypeChecked,
      // Alternatively, use this for stricter rules
      tseslint.configs.strictTypeChecked,
      // Optionally, add this for stylistic rules
      tseslint.configs.stylisticTypeChecked,

      // Other configs...

…

View on GitHub

How I Used Notion MCP

The true magic behind the Glassboard is its autonomous backend. Rather than hardcoding the 180+ projects or connecting directly to a rate-limited front-end API, I used the Notion MCP (Model Context Protocol) to turn my actual Notion Workspace into a headless CMS and evaluation engine:

The Living Database--> https://khanstruct.notion.site/notion-mcp-challenge?source=copy_link
Our central source of truth is The Notion AI Challenge Database, which holds every DEV.to submission, reaction stat, and read-time metric.

Hourly Autonomous Generation: Every single hour, our AI agent spins up via the local Notion MCP to read the latest rows from that master Notion database. It dynamically queries for any new entries or updated stats, then instantly builds and pushes those individual glassmorphism React cards directly into the Glassboard UI.

Two-Way Evaluation Sync: When a judge reviews a project and marks it with an implementation score (Originality, Practicality, Complexity), that data safely hits a securityHumanInLoop: true verification flag and writes safely back into the exact same Notion Database via MCP—enabling full remote collaboration for multiple judges without ever showing them a boring spreadsheet.

Notion + Twilio + WhatsApp = Give me what I want now

Zain K. — Mon, 30 Mar 2026 04:24:09 +0000

Cortana: WhatsApp Command Center for a Notion Workspace

What I Built

Cortana is a WhatsApp-to-Notion command center. You text a message from your phone, and Cortana:

Classifies the intent into one of 6 life domains (Blueprint)
Searches the matching knowledge in my Notion workspace
Replies back in WhatsApp with the domain + relevant pages

The 6 Blueprint domains:

Personal — identity, contacts, legal, finance, assets
Health — medical, fitness, medications, providers
Projects — code, hackathons, creative work
Work — employers, applications, consulting
Growth — learning, writing, AI exploration
Systems — archives, exports, logs, normalization

How It Works

flowchart LR
    A[WhatsApp message] --> B[Twilio webhook]
    B --> C[Cloudflare Worker adapter]
    C --> D[Notion Worker: classifyAndRespond]
    D --> E[Notion API search]
    E --> F[Formatted WhatsApp reply]

Request flow:

WhatsApp message arrives in Twilio
Twilio calls my Cloudflare Worker webhook
The adapter forwards into a Notion Worker tool (classifyAndRespond)
The tool classifies the message into a Blueprint domain and runs a Notion workspace search
The results are formatted and returned to Twilio
Twilio replies back to WhatsApp

The Stack

Notion Workers (alpha) — worker tool runtime
Cloudflare Workers — Twilio webhook adapter
Twilio WhatsApp API — message ingestion and reply
Notion API (@notionhq/client) — workspace search
TypeScript — end-to-end

How I Used Notion MCP

Notion is the system of record for this project:

My workspace is organized into the Blueprint taxonomy (Personal, Health, Projects, Work, Growth, Systems)
Cortana uses that structure to classify and retrieve the right information instead of doing a generic search
Notion AI was used to structure planning artifacts and build documentation during development

Demo

A WhatsApp question
The returned domain classification
A few matched Notion pages

The Build Story (key blocker + fix)

I hit a repeated deploy failure while using the Notion Workers CLI:

error: Failed to create worker: unauthorized
hint: run ntn login to authenticate

Login succeeded, but deploy kept failing. Root cause: a macOS Keychain issue where the CLI stored tokens but couldn’t reliably read them during deploy.

Fix:

export NOTION_API_TOKEN=$(security find-generic-password -s "notion-cli" -a "WORKSPACE_ID" -w)
export NOTION_WORKSPACE_ID=WORKSPACE_ID
ntn workers deploy

After that: Notion Worker deployed, Cloudflare adapter deployed, Twilio wired up, and messages round-tripped end-to-end.

What’s Next

Domain-filtered search (restrict search to a specific Blueprint domain)
triageBacklog (inbox triage and prioritization)
dispatchBlueprint (create new pages/rows directly into the right domain)
Multi-turn context (track conversation threads)
Move off Twilio sandbox to a real WhatsApp number

Notion + Twilio + WhatsApp = Give me what I want now

Zain K. — Mon, 30 Mar 2026 02:46:24 +0000

## Cortana: I Built an AI Command Center in One Night

TL;DR: I turned WhatsApp into a personal assistant that can search my Notion workspace and return answers instantly.

The Idea

I wanted something simple:

Instead of digging through Notion, opening tabs, and searching manually…
I could just text:

“Lab results from January”

…and get the answer back instantly.

What I Built

A lightweight system that connects:

WhatsApp → send a message
AI → figures out what you mean
Notion → searches your data
WhatsApp → sends results back

All in about 2 seconds.

It also organizes everything into 6 areas of life:

Personal
Health
Work
Projects
Growth
Data

What Went Wrong

This was supposed to take 2 hours. It took 6.

Three things broke:

Login/auth randomly failed (spent way too long on this)
The system I planned literally couldn’t work the way I designed it
Messages wouldn’t send because of a tiny formatting issue

Each fix was simple… once I figured it out.

What Works Now

You can text things like:

“Lab results January”
“Tell me about R-Cubed”
“What projects did I ship?”

And it replies with relevant results from Notion almost instantly.

No tabs. No searching. Just text → answer.

Why This Matters

This isn’t about WhatsApp.

It’s about removing friction between:

what you want
and where your knowledge lives

We’re getting closer to a world where your data is actually usable in real-time.

What’s Next

Smarter search (not just keywords)
Better conversations (follow-up questions)
Multi-user support
Full AI assistant experience

Code

https://github.com/zainkhan1994/cortana-notion-mcp

The Real Moment

At 4:30 AM I texted:

“Lab results January”

Got a reply in 2 seconds with exactly what I needed.

That’s when it clicked.

This is how interacting with your data should feel.

2am Debugger. 9am Builder.

Zain K. — Fri, 20 Feb 2026 01:38:18 +0000

AI systems thinker focused on automation, structured data, and scalable workflows.

DEV Community: Zain K.

Notion MCP Challenge Judges Dashboard

What I Built

DEMO

Show us the code

zainkhan1994 / notion-judge-glassboard

React + TypeScript + Vite

React Compiler

Expanding the ESLint configuration

How I Used Notion MCP

Notion + Twilio + WhatsApp = Give me what I want now

Cortana: WhatsApp Command Center for a Notion Workspace

What I Built

How It Works

The Stack

How I Used Notion MCP

Demo

The Build Story (key blocker + fix)

What’s Next

Links

Notion + Twilio + WhatsApp = Give me what I want now

The Idea

What I Built

What Went Wrong

What Works Now

Why This Matters

What’s Next

Code

The Real Moment

2am Debugger. 9am Builder.