MilonAI - Collaborative AI Chat Application

Landing Page
Dashboard Page
Chat Session

Inspiration

Artificial Intelligence is like a super-smart friend who can dig up answers, spark ideas, and tackle problems faster than we ever could alone. It’s great at crunching data, spotting patterns, and giving us a starting point for our work. But let’s be real—AI on its own is just one piece of the puzzle. Humans are messy, creative, and full of unique perspectives that no machine can match. When we work together, we bring different experiences, questions, and ideas to the table, making something bigger than any one of us could dream up.

Collaboration is how humans have always moved forward. Think about it: from building ancient cities to launching rockets to the moon, teamwork makes the impossible possible. It lets us challenge each other, fill in the gaps, and come up with ideas that are smarter, bolder, and more creative. Now, add AI to that mix, and you’ve got something special. AI doesn’t just answer questions—it listens to the group, builds on everyone’s input, and gives back ideas that are sharper because of it. Together, AI and human teamwork create a powerhouse: you get faster solutions, deeper insights, and more creative results than either could do alone. It’s like a brainstorming session where everyone, including the AI, brings their A-game.

Here why this combo rocks:

Bigger Ideas: Different people + AI = ideas that are more creative and well-rounded.
Smarter Solutions: Teamwork spots holes in plans, and AI fills them with facts and insights.
Faster Progress: Working together with AI cuts through the noise, getting you to answers quicker.
Shared Wins: Everyone learns from each other, making the whole group stronger.

To achieve this, we need a platform with features like:

✅ Create and manage multiple chat sessions with AI-powered responses.
✅ Select previous messages as context for more accurate AI replies.
✅ Public chat sharing via link with real-time collaboration and usernames displayed.
✅ Private chat sharing via email with controlled access and contributions.
✅ Real-time updates and typing indicators using WebSockets.

Example of real-world scenarios where collaborative AI can help.

In the Classroom
A teacher’s lecture gives you the basics, but it’s the students’ questions that really bring a topic to life. One kid might ask about the science of something, another about its history, and suddenly everyone’s learning more than they would’ve alone. A regular AI chat app? It’s just you and the AI, so you miss out on what others might ask. The knowledge stays stuck with one person. Collaborative AI flips this. Imagine a biology class studying ecosystems. Students hop into a public chat with a shared link, each tossing in questions like “How do coral reefs help ocean life?” or “What do predators do for balance?” The AI answers one student, but everyone sees it and learns from it. One kid’s question about climate change sparks another about fixing damaged reefs, building a bigger picture together. The teacher can join the chat to guide things or add better questions. Unlike a solo AI app, collaborative AI makes every question a group win, helping the whole class understand the topic better.

In a Startup
A group tackles a problem designing eco-friendly packaging, and everyone sees something different. An engineer thinks about materials, a marketer about what customers want, and an environmentalist about the planet. A regular AI chat app can’t keep up—it’s just one person asking questions, so the AI’s answers miss the bigger picture.

With Collaborative AI, the team starts a private chat, inviting only the right people by email. The engineer asks the AI about sustainable materials, building on a past answer about recycling. The AI gives a detailed response, and the marketer jumps in with a question about what customers will love. The environmentalist asks about pollution, and the AI tweaks its answers based on everyone’s input. The team watches ideas come together in real time. Unlike a regular AI app, MilonAI feels like a group huddle where the AI is a teammate, helping build solutions that are smarter and more complete.

What it does

MilonAI is a collaborative AI chat platform designed to combine the intelligence of artificial intelligence with the creativity and curiosity of human teams. Unlike traditional AI chat apps where one person interacts with the AI in isolation, MilonAI allows multiple users to interact with the AI together—making learning, problem-solving, and brainstorming a shared experience.

🚀 Key Features

Multi-User Chat Sessions
Start a conversation with the AI and invite others to join, either publicly via a shareable link or privately via email.
Context-Aware Responses
Users can select past messages to provide context, improving the relevance and depth of AI replies.
Public & Private Collaboration
Share conversations with the world or keep them private for your team, classroom, or group.
Real-Time Interaction
Built with WebSockets, MilonAI supports live typing indicators, user name tags, and dynamic updates so everyone stays in sync.

How we built it

MilonAI was built using modern web technologies with the assistance of the Bolt.new AI development tool. Below is a breakdown of the key technical components and highlights that power the platform.

Tech Stack Overview:

Frontend: React + TailwindCSS
Backend: Node.js with Express
Database: PostgreSQL (using Knex for queries)
Authentication: JSON Web Tokens (JWT)
AI Integration: Google AI API (manually configured via API key)
Real-time Collaboration: WebSockets (via Socket.IO)
Deployment: Render

Key Technical Highlights:

Message context management: Users can select which past messages to include in context, giving them control over AI memory without overwhelming the model.
Public/private sharing: Public chats generate sharable URLs; private ones require email invitations and enforce access control.
Live feedback: WebSockets provide typing indicators, name tags, and real-time updates to simulate an organic team chat.

Challenges we ran into

Context Limitation: Handling which messages should be sent as context to the AI without exceeding token limits required a smart selection mechanism and UI clarity.
Real-Time Sync: WebSockets are powerful but also tricky when ensuring all users see the same updates in the right order.
User Roles & Access Control: Public vs. private collaboration demanded strict access logic and a flexible sharing system.
AI Responsiveness: Making the AI feel “collaborative” meant tuning prompts and responses based on shared user input—trial and error helped refine the tone and relevance.
Hosting Constraints: Building real-time features on a limited budget forced me to optimize for performance and database queries efficiently.

Accomplishments that we're proud of

Building MilonAI was an exciting journey filled with challenges, creativity, and technical growth. Here are some key accomplishments we're especially proud of:

✅ Designed a Real-Time Collaborative AI Chat Experience
Created a platform where multiple users can interact with an AI simultaneously—something no AI chat tools offer.
✅ Seamless Context Management for Smarter AI Replies
Implemented a message selection system that allows users to control what the AI remembers, resulting in more accurate and coherent responses.
✅ Public and Private Chat Sharing with Live Participation
Developed a flexible sharing model where conversations can be public (via link) or private (via secure email invite), enabling collaboration in both open and restricted settings.
✅ Integrated WebSockets for Real-Time Typing Indicators and Updates that I have never done before
Made the chat experience feel dynamic and human by adding live typing status, user tags, and instant updates across participants.
✅ Built a Full-Stack App from Scratch with Modern Technologies, most of which wehad never used before
Used a powerful combination of React, Node.js, Express, PostgreSQL, WebSockets, and TailwindCSS—deployed on Render—to bring the entire idea to life.
✅ Inspired Use in Real-World Contexts
From classrooms to startups, MilonAI can be a very useful tool for group learning, idea sharing, and decision-making in collaborative environments.

These accomplishments not only reflect what we built—but also what’s possible when humans and AI work together as creative partners.

What we learned

MilonAI wasn’t just a coding exercise—it was a study in human-AI collaboration.

I learned how shared context improves AI quality, and how to engineer that into a multi-user system.
I gained deeper insight into real-time communication models using WebSockets and how to maintain state across multiple clients.
I explored UI/UX patterns for conversational apps that encourage collaboration rather than isolation.
I understood the balance between user control and AI-driven intelligence—how much should users guide the AI, and how much should the AI infer?
I learned about new technologies Node.js, Express, and WebSockets.

Building MilonAI showed me that AI’s power isn’t just in answers—it’s in how it helps people build better ideas together.

What's next for MilonAI - Collaborative AI Chat Application

While MilonAI already offers a powerful and collaborative experience, we’re just getting started. Here's a glimpse of what we’re working on next:

✨ Upcoming Features

🖼️ Image Prompts
Upload images to use as visual prompts—perfect for analyzing product labels, diagrams, or creative inspiration.
🎙️ Voice Prompts
Talk to the AI directly with real-time voice input, making the experience more natural and hands-free.
🧠 Image Responses
Receive AI-generated images (e.g., charts, sketches, visual explanations) directly in the chat for more expressive replies.
🔊 Voice Responses
Get audio replies from the AI—great for accessibility and multitasking scenarios.

🛠️ Purpose-Built Tools

We plan to expand MilonAI into a suite of domain-specific tools, each tailored to real-world needs. These tools will be powered by the same collaborative AI core, but fine-tuned for specific use cases:

👩‍🏫 AI Teacher
A guided learning assistant for classrooms and tutoring, supporting real-time group questions, lesson summaries, and study aids.
💬 FAQ Bot for Websites
An embeddable AI assistant that provides dynamic, contextual answers for customers—learning from website content and public customer chat rooms.