AIdeas: SenseSync - The Third Channel
For 187 years, digital communication has wired up exactly two of the five human senses. SenseSync opens the third. A real-time haptic communication app built on AWS IoT Core, Lambda, Bedrock, and DynamoDB that lets people send rhythm and presence across any distance. Zero dollars a month. The haptic channel, finally online.
App Category
Social Impact
My Vision
The telegraph was invented in 1837.
Within a decade, humans could send text across continents instantly. Within another decade, we sent voice. A century later, video. Each time, we took something that only worked in person and made it work at any distance.
We have been doing this for 187 years.
And in all that time, we only ever wired up two of the five senses.
- 👀 Vision — digitized (video, images, text)
- 👂 Hearing — digitized (calls, voice messages)
- 👋 Touch — never digitized
- 👃 Smell — not yet
- 🍽️ Taste — not yet
Smell and taste are genuinely hard. The hardware doesn't exist in everyone's pocket.
But touch? The hardware has been sitting in every smartphone for over a decade. The vibration motor. The haptic engine. The same chip that buzzes when you get a notification has been waiting, completely unused as a communication channel, since the iPhone shipped in 2007.
Nobody asked what else it could say.
SenseSync is the answer to that question.
Not a simulation of touch. Not fake hugs through a screen. Something more interesting than that: a new communication channel, built for the body, that has never existed before.
Radio didn't try to replicate the sound of a person standing next to you. It created something new: broadcasting. SenseSync doesn't try to replicate physical touch. It creates something new: haptic communication. A language of rhythm, intensity, and timing that two people can share across any distance, felt directly by the body, understood without words.
1
You tap → AWS transmits → They feel it
Five patterns ship with the app, each one a word in this new language:
- 💓 Heartbeat — I am thinking of you
- 🌊 Wave — Hello from wherever I am
- 🤗 Hug — One long, steady presence
- ✨ Gentle Pulse — I am here, no need to reply
- 🚨 Urgent Alert — I need you, right now
And for everything that those five words cannot say, SenseSync has an AI-powered Pattern Creator. Describe what you want to communicate in plain language. Amazon Bedrock with Claude converts that description into a haptic pattern tuned to its emotional shape. A new word, invented by you, felt by them.
Built for the AWS 10,000 AIdeas Competition using Kiro and the AWS Free Tier. Monthly infrastructure cost: $0.
Why This Matters
Every communication technology we have built rewires the same two channels: eyes and ears.
Video calls, podcasts, social media, live streaming, VR headsets. All of it lands on the visual or auditory cortex. The haptic cortex, which processes touch and physical sensation, has received exactly nothing from the digital world in 187 years of telecommunications history.
This is not a small gap. Touch is the first sense that develops in the womb. It is the sense that never fully sleeps, even when we do. Neuroscientists describe it as the sense most tightly coupled to emotional regulation, social bonding, and the basic feeling of not being alone.
We built the internet. We gave it two senses. We left one out.
SenseSync opens the third channel. Here is who it matters to first:
💑 Long-Distance Relationships and Separated Families
Words are interpretable. Tone can be misread. But a steady 800ms vibration at 2am carries a meaning that needs no translation: I am awake, I am thinking of you, you are not alone. No call needed. No text composed. One tap. They feel it.
🤟 Deaf and Hard-of-Hearing Users
For users who are deaf or hard-of-hearing, SenseSync opens a communication channel that bypasses sound entirely and lands directly on the body. A vocabulary of rhythm and intensity that is entirely tactile. This is not an accessibility workaround. It is a native channel for a community that has always communicated through physical sensation, now extended across distance for the first time.
🎮 Gaming and Shared Experiences
The same real-time haptic delivery infrastructure that sends a heartbeat can synchronize physical sensations between players, sync vibrations to live audio, or create shared physical moments during remote co-op experiences. The channel SenseSync built for human connection has applications anywhere two people want to feel something together from different locations.
How I Built This
The Core Engineering Problem
Haptic communication fails if it is late. A heartbeat pattern that arrives 3 seconds after it was sent is not a heartbeat. It is a notification. The entire design of SenseSync's backend was built around one constraint: delivery in under 100 milliseconds, at zero infrastructure cost.
That constraint drove every architecture decision.
Architecture
AWS Services
| Service | Role |
|---|---|
| Amazon Cognito | Phone number auth and JWT tokens |
| AWS API Gateway | REST API and WebSocket endpoint |
| AWS Lambda | sendTouch, receiveTouch, generatePattern, manageContacts |
| Amazon DynamoDB | Users, contacts, touch history, custom patterns |
| AWS IoT Core | Real-time MQTT delivery, the core of the system |
| Amazon SNS | Push notifications via APNs (iOS) and FCM (Android) |
| Amazon Bedrock | AI haptic pattern generation with Claude |
| Amazon CloudWatch | Latency monitoring and billing alerts |
| Amazon S3 + CloudFront | Static assets and global CDN |
Total monthly cost: $0 ✅
Designing the Language
The hardest part of SenseSync was not the infrastructure. It was the patterns themselves.
A haptic pattern is defined by three variables per step: duration (how long the vibration lasts), intensity (light, medium, or heavy), and pause (silence before the next step). The heartbeat pattern has four steps. It took 12 iterations to get right.
1
2
3
4
5
6
7
8
9
10
11
export const builtInPatterns = {
heartbeat: [
{ duration: 80, intensity: 'light', pause: 80 },
{ duration: 120, intensity: 'heavy', pause: 400 },
{ duration: 80, intensity: 'light', pause: 80 },
{ duration: 120, intensity: 'heavy', pause: 400 },
],
hug: [
{ duration: 800, intensity: 'medium', pause: 0 },
],
};
The second beat of the heartbeat needs to be heavier than the first. The pause after the pair needs to be long enough to feel like breathing. Move any value by 20ms in the wrong direction and it stops feeling like a heartbeat and starts feeling like a glitch. Every millisecond is a design decision.
Real-Time Delivery via AWS IoT Core
Each user gets a private MQTT topic on AWS IoT Core:
touches/{userId}. When a touch is sent, Lambda publishes to that topic instantly. The recipient's app, subscribed via AWS Amplify PubSub, receives it and plays the pattern:1
2
3
4
5
PubSub.subscribe(`touches/${userId}`).subscribe({
next: async (data) => {
await playPattern(data.value.steps);
}
});
If the recipient's app is closed, Lambda falls back automatically to Amazon SNS, which routes to APNs (iPhone) or FCM (Android) and wakes the device within roughly 2 seconds. No touch is ever dropped.
AI Pattern Generation
The Pattern Creator sends a natural language description to Amazon Bedrock with a prompt engineered to produce consistent, structured haptic output:
1
2
3
4
5
6
7
8
The user wants to communicate: "warm and slow, like a Sunday morning"
You are a haptic pattern designer for a communication app.
Return ONLY valid JSON with a maximum of 6 steps.
Each step: { duration: 50-800ms, intensity: light|medium|heavy, pause: 50-500ms }
light = gentle, quiet medium = warm, present heavy = direct, urgent
Match the emotional character of the description to the rhythm and weight of the pattern.
The output is a typed JSON pattern, saved to DynamoDB, previewable on a live waveform visualizer before sending. A feeling described in words becomes a feeling transmitted through vibration.
Development Timeline
- Week 1 — Cognito auth with phone and OTP, navigation, foundational architecture
- Week 2 — Haptic engine, 5 core patterns, contacts system, DynamoDB schema
- Week 3 — IoT Core real-time delivery, SNS push notification fallback
- Week 4 — Bedrock AI pattern creator, touch history, waveform visualizer, performance tuning
Performance
| Metric | Target | Result |
|---|---|---|
| Touch delivery (app open) | < 100ms | 87ms ✅ |
| Touch delivery (app closed) | < 5s | ~2s ✅ |
| AI pattern generation | < 3s | ~1.4s ✅ |
| API response time | < 200ms | 145ms ✅ |
| Monthly AWS cost | $0 | $0 ✅ |
Demo
Home Screen
Select a contact, choose a pattern, tap. Your contact feels it in real time. The UI pulses with the rhythm of your selected pattern as a preview before you send.
📸 [Screenshot: Home screen with heartbeat selected and contact shown]
AI Pattern Creator
Type a feeling. A waveform appears. Preview it on your own phone first. Send it.
📸 [Screenshot: Pattern Creator with text input and generated waveform]
Touch History
Every touch sent and received, with delivery method, latency, and pattern type. Tap any entry to feel it again.
📸 [Screenshot: History screen with touch timeline]
Contacts
Search by username, send a request, connect. Touch flows both ways once accepted.
📸 [Screenshot: Contacts screen with connected users]
What I Learned
1. The third channel was always there. Nobody asked what it could say.
The haptic motor in a smartphone has existed since the early 2000s. For two decades it has been used for exactly one thing: notifications. SenseSync didn't invent new hardware. It asked a different question about hardware that already existed. Sometimes the most important innovations are not about building something new but about looking at something old and asking what else it could be.
2. Designing in milliseconds teaches you that emotion lives in timing.
Writing code is largely forgiving. Being 20ms off in a heartbeat pattern is not. The difference between a pattern that feels human and one that feels mechanical is often a single timing value. This forced a level of precision I had never applied to product design before. Haptic design is closer to music composition than software engineering.
3. A better prompt beats a bigger model every time.
Early versions of the AI pattern generator produced creative but unparseable output. The fix was not switching models. It was treating the prompt like a typed schema contract: explicit field names, numeric ranges, intensity definitions, and emotional mappings spelled out in plain language. Same model, same data, completely different quality. The prompt is the product.
4. Kiro removed the ceiling on what one person can build.
The Lambda functions, IoT routing, DynamoDB schema, and SAM template that would have taken days to wire up manually were scaffolded by Kiro in hours. That time went back into the product: into the timing of the heartbeat, the waveform visualizer, the feel of the AI output. The best infrastructure is the kind you never have to think about.
5. Building at zero cost is a statement about who can use what you make.
Every architecture decision that kept SenseSync inside the AWS Free Tier also made it deployable by any developer anywhere, at no cost. For a communication tool designed to connect people across distance, that matters. Accessibility is not just a feature. It is a distribution strategy.
Radio gave the world a new way to send sound.
SenseSync gives the world a new way to send presence.
187 years after the telegraph. The third channel is open.
Built for the AWS 10,000 AIdeas Competition — Social Impact Category: SenseSync | Harish Muthyala | mharishchowdary11@gmail.com
Enjoyed reading this content? Let the author know!
Your likes, comments, shares, and saves help creators reach more builders.
Loading recommendations
Loading article